WebKit-https.git
3 months agoExpose full screen controller SPI to check if full screen is open and close full...
commit-queue@webkit.org [Fri, 3 May 2019 17:34:46 +0000 (17:34 +0000)]
Expose full screen controller SPI to check if full screen is open and close full screen
https://bugs.webkit.org/show_bug.cgi?id=197263

Patch by Jay Mulani <jmulani@apple.com> on 2019-05-03
Reviewed by Jer Noble.

* UIProcess/API/Cocoa/WKWebView.mm:
(-[WKWebView closeFullScreenWindowController]):
(-[WKWebView fullScreenWindowController]):
* UIProcess/API/Cocoa/WKWebViewInternal.h:
* UIProcess/API/Cocoa/WKWebViewPrivate.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244911 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoLayoutTest imported/w3c/web-platform-tests/xhr/event-upload-progress-crossorigin...
youenn@apple.com [Fri, 3 May 2019 17:32:25 +0000 (17:32 +0000)]
LayoutTest imported/w3c/web-platform-tests/xhr/event-upload-progress-crossorigin.htm is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=182849
<rdar://problem/50353372>

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

Modernize the test and add an assertion that the body is transmitted after a redirection.

* web-platform-tests/xhr/event-upload-progress-crossorigin.htm:

Source/WebKit:

In case of restarting the load after a redirection, which now happens in case
of following cross origin redirections with same-origin credentials, make sure to
update the request as would do NetworkLoad.
This is in particular important to preserve the request body.

Covered by WPT updated test.

* NetworkProcess/NetworkLoad.cpp:
(WebKit::updateRequest):
(WebKit::NetworkLoad::updateRequestAfterRedirection const):
(WebKit::NetworkLoad::continueWillSendRequest):
* NetworkProcess/NetworkLoad.h:
* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::continueWillSendRequest):

LayoutTests:

* platform/mac-wk2/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244910 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoDo not require log_to_stderr for WebRTC logging through WebKit
youenn@apple.com [Fri, 3 May 2019 16:39:05 +0000 (16:39 +0000)]
Do not require log_to_stderr for WebRTC logging through WebKit
https://bugs.webkit.org/show_bug.cgi?id=197560

Reviewed by Eric Carlson.

* Source/webrtc/rtc_base/logging.cc:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244909 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: Record actions performed on WebGL2RenderingContext
drousso@apple.com [Fri, 3 May 2019 16:36:17 +0000 (16:36 +0000)]
Web Inspector: Record actions performed on WebGL2RenderingContext
https://bugs.webkit.org/show_bug.cgi?id=176008
<rdar://problem/34213884>

Reviewed by Joseph Pecoraro.

Source/JavaScriptCore:

* inspector/protocol/Recording.json:
* inspector/scripts/codegen/generator.py:
Add `canvas-webgl2` as a `Type`.

Source/WebCore:

Tests: inspector/canvas/recording-webgl2.html
       inspector/canvas/recording-webgl2-snapshots.html

* html/canvas/WebGL2RenderingContext.idl:

* bindings/js/CallTracerTypes.h:
* inspector/RecordingSwizzleTypes.h:

* inspector/InspectorCanvas.h:
* inspector/InspectorCanvas.cpp:
(WebCore::InspectorCanvas::canvasChanged): Added.
(WebCore::InspectorCanvas::resetRecordingData):
(WebCore::shouldSnapshotWebGL2Action): Added.
(WebCore::InspectorCanvas::recordAction):
(WebCore::InspectorCanvas::releaseObjectForRecording):
(WebCore::InspectorCanvas::appendActionSnapshotIfNeeded):
(WebCore::InspectorCanvas::buildAction):
* inspector/agents/InspectorCanvasAgent.h:
* inspector/agents/InspectorCanvasAgent.cpp:
(WebCore::InspectorCanvasAgent::canvasChanged): Added.
(WebCore::InspectorCanvasAgent::startRecording):
Provide an actual implementation of `CanvasObserver::canvasChanged` since it can be used to
determine whether or not an action needs a snapshot.

* page/PageConsoleClient.cpp:
(WebCore::canvasRenderingContext):

Source/WebInspectorUI:

* UserInterface/Models/Recording.js:
(WI.Recording.fromPayload):
(WI.Recording.displayNameForSwizzleType):
(WI.Recording.prototype.async swizzle):
(WI.Recording.prototype.createContext):

* UserInterface/Models/RecordingAction.js:
(WI.RecordingAction.constantNameForParameter):
(WI.RecordingAction._prototypeForType):
(WI.RecordingAction.prototype.async swizzle):
Treat all actions with a snapshot as being a "visible" action.

* UserInterface/Views/CanvasTabContentView.css:
(.content-view.tab.canvas .navigation-bar > .item .canvas:matches(.webgl, .webgl2, .webgpu, .webmetal) .icon): Added.
(.content-view.tab.canvas .navigation-bar > .item .canvas.webgl): Deleted.
* UserInterface/Views/CanvasContentView.js:
(WI.CanvasContentView.prototype.initialLayout):
* UserInterface/Views/CanvasSidebarPanel.js:
(WI.CanvasSidebarPanel.prototype.shown):
(WI.CanvasSidebarPanel.prototype._canvasChanged):
(WI.CanvasSidebarPanel.prototype._updateRecordNavigationItem):

* UserInterface/Views/RecordingContentView.js:
(WI.RecordingContentView):
(WI.RecordingContentView.prototype.get navigationItems):
(WI.RecordingContentView.prototype.shown):
(WI.RecordingContentView.prototype.initialLayout):
* UserInterface/Views/RecordingActionTreeElement.js:
(WI.RecordingActionTreeElement._generateDOM.createParameterElement):

LayoutTests:

* inspector/canvas/recording-webgl2.html: Added.
* inspector/canvas/recording-webgl2-expected.txt: Added.
* inspector/canvas/recording-webgl2-snapshots.html: Added.
* inspector/canvas/recording-webgl2-snapshots-expected.txt: Added.

* TestExpectations:
* platform/gtk/TestExpectations:
* platform/mac/TestExpectations:
* platform/win/TestExpectations:
* platform/wincairo/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244908 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244881.
commit-queue@webkit.org [Fri, 3 May 2019 14:32:23 +0000 (14:32 +0000)]
Unreviewed, rolling out r244881.
https://bugs.webkit.org/show_bug.cgi?id=197559

Breaks compilation of jsconly on linux, breaking compilation
for jsc-i386-ews, jsc-mips-ews and jsc-armv7-ews (Requested by
guijemont on #webkit).

Reverted changeset:

"[CMake] Refactor WEBKIT_MAKE_FORWARDING_HEADERS into
WEBKIT_COPY_FILES"
https://bugs.webkit.org/show_bug.cgi?id=197174
https://trac.webkit.org/changeset/244881

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244907 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoResolve the percentage values of inset properties against proper box.
commit-queue@webkit.org [Fri, 3 May 2019 13:19:57 +0000 (13:19 +0000)]
Resolve the percentage values of inset properties against proper box.
https://bugs.webkit.org/show_bug.cgi?id=189549

Patch by Joonghun Park <jh718.park@samsung.com> on 2019-05-03
Reviewed by Antti Koivisto.

Before this CL, sticky element's layout was executed relative to
a box's overflow container,
but the value returned by getComputedStyle was resolved against
its containing block.

So, the computed value and the actual value used in layout
was different before this change.

LayoutTests/imported/w3c:

* web-platform-tests/css/cssom/getComputedStyle-insets-sticky-container-for-abspos-expected.txt: Copied from LayoutTests/imported/w3c/web-platform-tests/css/cssom/getComputedStyle-insets-sticky-expected.txt.
* web-platform-tests/css/cssom/getComputedStyle-insets-sticky-container-for-abspos.html: Copied from LayoutTests/imported/w3c/web-platform-tests/css/cssom/getComputedStyle-insets-sticky.html.
* web-platform-tests/css/cssom/getComputedStyle-insets-sticky-expected.txt:
* web-platform-tests/css/cssom/getComputedStyle-insets-sticky.html:
* web-platform-tests/css/cssom/getComputedStyle-sticky-pos-percent-expected.txt: Added.
* web-platform-tests/css/cssom/getComputedStyle-sticky-pos-percent.html: Added.
* web-platform-tests/css/cssom/support/getComputedStyle-insets.js:
(runTestsWithWM):

Source/WebCore:

Tests: imported/w3c/web-platform-tests/css/cssom/getComputedStyle-insets-sticky-container-for-abspos.html
       imported/w3c/web-platform-tests/css/cssom/getComputedStyle-sticky-pos-percent.html

* css/CSSComputedStyleDeclaration.cpp:
(WebCore::positionOffsetValue):
* rendering/RenderBox.cpp:
(WebCore::RenderBox::enclosingScrollportBox const):
* rendering/RenderBox.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244906 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed WPE build fix after r244898.
zandobersek@gmail.com [Fri, 3 May 2019 08:42:18 +0000 (08:42 +0000)]
Unreviewed WPE build fix after r244898.

* WebKitTestRunner/wpe/TestControllerWPE.cpp:
(WTR::TestController::abortModal): Add the missing (no-op) definition.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244905 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd a quirk to make gmail navigation bar scrollable without mouse hover on iOS
antti@apple.com [Fri, 3 May 2019 06:48:14 +0000 (06:48 +0000)]
Add a quirk to make gmail navigation bar scrollable without mouse hover on iOS
https://bugs.webkit.org/show_bug.cgi?id=197529
<rdar://problem/49403416>

Reviewed by Simon Fraser.

* css/StyleResolver.cpp:
(WebCore::StyleResolver::adjustRenderStyle):
(WebCore::StyleResolver::adjustRenderStyleForSiteSpecificQuirks):

Turn 'overflow:hidden' to 'overflow:auto' on element with role="navigation".
This should be both reasonably targeted and robust.

* css/StyleResolver.h:
* page/Quirks.cpp:
(WebCore::Quirks::needsGMailOverflowScrollQuirk const):
* page/Quirks.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244904 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMark U+01C0 as a homograph of U+006C
commit-queue@webkit.org [Fri, 3 May 2019 06:23:59 +0000 (06:23 +0000)]
Mark U+01C0 as a homograph of U+006C
https://bugs.webkit.org/show_bug.cgi?id=197526
<rdar://problem/50301904>

Patch by Alex Christensen <achristensen@webkit.org> on 2019-05-02
Reviewed by Tim Horton.

Source/WTF:

* wtf/URLHelpers.cpp:
(WTF::URLHelpers::isLookalikeCharacter):

Tools:

* TestWebKitAPI/Tests/WTF/cocoa/URLExtras.mm:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244903 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Inspector: Dark Mode: borders between section should be lighter than background
nvasilyev@apple.com [Fri, 3 May 2019 06:17:09 +0000 (06:17 +0000)]
Web Inspector: Dark Mode: borders between section should be lighter than background
https://bugs.webkit.org/show_bug.cgi?id=197551

Reviewed by Joseph Pecoraro.

* UserInterface/Views/DetailsSection.css:
(.details-section .details-section):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244902 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMake focusing-element-with-tabindex-by-tap-or-click.html more robust on iOS
rniwa@webkit.org [Fri, 3 May 2019 05:02:30 +0000 (05:02 +0000)]
Make focusing-element-with-tabindex-by-tap-or-click.html more robust on iOS
https://bugs.webkit.org/show_bug.cgi?id=197548

Reviewed by Wenson Hsieh.

There was a race between the time the editable element was focused and the keyboard was brought up
to when the next non-editable element was focused. Made the test explicitly wait for the keyboard
to come up and go down between test cases to make it more robust.

* fast/events/focusing-element-with-tabindex-by-tap-or-click.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244901 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd logging for RenderLayer clip rects
simon.fraser@apple.com [Fri, 3 May 2019 05:01:49 +0000 (05:01 +0000)]
Add logging for RenderLayer clip rects
https://bugs.webkit.org/show_bug.cgi?id=197547

Reviewed by Zalan Bujtas.

Add a ClipRects log channel, and stream output for ClipRect and ClipRects.

* platform/Logging.h:
* rendering/ClipRect.cpp:
(WebCore::operator<<):
* rendering/ClipRect.h:
* rendering/RenderLayer.cpp:
(WebCore::operator<<):
(WebCore::RenderLayer::calculateClipRects const):
* rendering/RenderLayer.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244900 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMake AudioContext::scriptExecutionContext() private
youenn@apple.com [Fri, 3 May 2019 04:28:13 +0000 (04:28 +0000)]
Make AudioContext::scriptExecutionContext() private
https://bugs.webkit.org/show_bug.cgi?id=197512

Reviewed by Eric Carlson.

Refactor code to make audio nodes not rely on AudioContext::scriptExecutionContext.
Instead, let AudioContext provide the necessary API for its nodes.
Covered by existing tests.

* Modules/webaudio/AudioBufferSourceNode.cpp:
(WebCore::AudioBufferSourceNode::looping):
(WebCore::AudioBufferSourceNode::setLooping):
* Modules/webaudio/AudioContext.cpp:
(WebCore::AudioContext::postTask):
(WebCore::AudioContext::origin const):
(WebCore::AudioContext::addConsoleMessage):
* Modules/webaudio/AudioContext.h:
(WebCore::AudioContext::isStopped const):
* Modules/webaudio/AudioNode.cpp:
(WebCore::AudioNode::scriptExecutionContext const):
* Modules/webaudio/AudioNode.h:
* Modules/webaudio/AudioScheduledSourceNode.cpp:
(WebCore::AudioScheduledSourceNode::finish):
* Modules/webaudio/DefaultAudioDestinationNode.cpp:
(WebCore::DefaultAudioDestinationNode::resume):
(WebCore::DefaultAudioDestinationNode::suspend):
(WebCore::DefaultAudioDestinationNode::close):
* Modules/webaudio/MediaElementAudioSourceNode.cpp:
(WebCore::MediaElementAudioSourceNode::wouldTaintOrigin):
* Modules/webaudio/MediaStreamAudioDestinationNode.cpp:
(WebCore::MediaStreamAudioDestinationNode::MediaStreamAudioDestinationNode):
* Modules/webaudio/ScriptProcessorNode.cpp:
(WebCore::ScriptProcessorNode::fireProcessEvent):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244899 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd test coverage for <rdar://problem/49731231>
cdumez@apple.com [Fri, 3 May 2019 01:19:03 +0000 (01:19 +0000)]
Add test coverage for <rdar://problem/49731231>
https://bugs.webkit.org/show_bug.cgi?id=196730
<rdar://problem/49731231>

Reviewed by Ryosuke Niwa.

Tools:

Add support for testRunner.abortModal() in macOS's WebKitTestRunner in order
to match DumpRenderTree.

* WebKitTestRunner/InjectedBundle/Bindings/TestRunner.idl:
* WebKitTestRunner/InjectedBundle/TestRunner.cpp:
(WTR::TestRunner::abortModal):
* WebKitTestRunner/InjectedBundle/TestRunner.h:
* WebKitTestRunner/TestController.cpp:
(WTR::TestController::runModal):
* WebKitTestRunner/TestController.h:
* WebKitTestRunner/TestInvocation.cpp:
(WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
* WebKitTestRunner/gtk/TestControllerGtk.cpp:
(WTR::TestController::abortModal):
* WebKitTestRunner/ios/TestControllerIOS.mm:
(WTR::TestController::abortModal):
* WebKitTestRunner/mac/TestControllerMac.mm:
(WTR::TestController::abortModal):
* WebKitTestRunner/win/TestControllerWin.cpp:
(WTR::TestController::abortModal):

LayoutTests:

* http/tests/security/showModalDialog-sync-cross-origin-page-load-expected.txt: Added.
* http/tests/security/showModalDialog-sync-cross-origin-page-load.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244898 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoREGRESSION: Layout test editing/selection/ios/selection-after-changing-text-with...
wenson_hsieh@apple.com [Fri, 3 May 2019 01:05:32 +0000 (01:05 +0000)]
REGRESSION: Layout test editing/selection/ios/selection-after-changing-text-with-callout-menu.html is failing
https://bugs.webkit.org/show_bug.cgi?id=197532
<rdar://problem/50177144>

Reviewed by Ryosuke Niwa.

Fixes layout tests that began failing after r244546. See below for details.

* UIProcess/RemoteLayerTree/RemoteLayerTreeDrawingAreaProxy.mm:
(WebKit::RemoteLayerTreeDrawingAreaProxy::commitLayerTree):

Partially reverts a change in r244546, after which we commit the layer tree and scroll before updating the
page's editor state. The purpose of this change was to ensure that UI process-side element focus scrolling logic
would not conflict with web-process-driven scrolling logic.

Instead, we split the existing logic in WebPageProxy::editorStateChanged into two pieces: one that updates the
editor state (by setting m_editorState), and a second that dispatches updates to PageClient when the first
editor state is received after focusing an element. During a layer tree commit in the UI process, we first
update the editor state, then commit the layer tree and apply scroll position changes, and finally scroll to
reveal the focused element if necessary.

When an editor state is delivered to the UI process in an out-of-band update (i.e. not in a layer tree commit),
simply dispatch the initial editor state for a focused element immediately.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::scheduleFullEditorStateUpdate):

Add an IPC message to schedule an editor state update in the next remote layer tree commit. See below for more
details.

(WebKit::WebPageProxy::editorStateChanged):
(WebKit::WebPageProxy::dispatchDidReceiveEditorStateAfterFocus):
* UIProcess/WebPageProxy.h:
* UIProcess/gtk/WebPageProxyGtk.cpp:
(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.
* UIProcess/ios/WKContentViewInteraction.mm:
(-[WKContentView willFinishIgnoringCalloutBarFadeAfterPerformingAction]):

Additionally ensure that an editor state update is scheduled. This addresses a potential source of flakiness in
the layout test editing/selection/ios/selection-after-changing-text-with-callout-menu.html, where an editor
state update may only be scheduled after the next layout timer fires (this is the case in custom callout menu
actions that change the DOM but do not otherwise trigger any editing commands).

In the problematic scenario, the client could make a change that triggers layout soon; but before the layout
timer fires, the timer for the next remote layer tree commit could fire, such that the next layer tree commit
would not contain the relevant editor state.

This extra step ensures that we always *schedule* an editor state update when performing a callout menu action
that does not automatically dismiss, so that we can prevent the callout bar from dismissing during the correct
scope.

* UIProcess/ios/WebPageProxyIOS.mm:
(WebKit::WebPageProxy::layerTreeCommitComplete):
(WebKit::WebPageProxy::updateEditorState):

Rename editorStateChanged to updateEditorState, and make the editorStateChanged codepath only executed when an
out-of-band editor state update is delivered to the UI process.

(WebKit::WebPageProxy::dispatchDidReceiveEditorStateAfterFocus):

Notify the UI process that the initial editor state has been received; this prompts us to scroll to reveal the
focused element, if needed.

(WebKit::WebPageProxy::editorStateChanged): Deleted.
* UIProcess/mac/WebPageProxyMac.mm:
(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.
* UIProcess/win/WebPageProxyWin.cpp:
(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.
* UIProcess/wpe/WebPageProxyWPE.cpp:
(WebKit::WebPageProxy::updateEditorState):
(WebKit::WebPageProxy::editorStateChanged): Deleted.
* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/WebPage.messages.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244897 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[CMake] Add support for LTO builds
chris.reid@sony.com [Thu, 2 May 2019 23:39:35 +0000 (23:39 +0000)]
[CMake] Add support for LTO builds
https://bugs.webkit.org/show_bug.cgi?id=188986

Reviewed by Don Olmstead.

.:

Add LTO_MODE cmake variable to enable flto in clang builds.
Add support to use ld.lld to build with LTO.

* Source/cmake/OptionsCommon.cmake:
* Source/cmake/WebKitCompilerFlags.cmake:

Tools:

Add support to configure --lto-mode with cmake builds.

* Scripts/build-webkit:
* Scripts/webkitdirs.pm:
(generateBuildSystemFromCMakeProject):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244895 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoDisable software keyboard for a math field textarea on desmos.com
rniwa@webkit.org [Thu, 2 May 2019 23:31:42 +0000 (23:31 +0000)]
Disable software keyboard for a math field textarea on desmos.com
https://bugs.webkit.org/show_bug.cgi?id=197488

Reviewed by Wenson Hsieh.

Treat a textarea inside a math field span as if it had inputmode content attribute set to none to suppress
the software keyboard on desmos.com as it interferes with website's own UI.

* html/HTMLElement.cpp:
(WebCore::HTMLElement::canonicalInputMode const):
* page/Quirks.cpp:
(WebCore::Quirks::needsInputModeNoneImplicitly const):
* page/Quirks.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244894 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoNSAttributedString conversion in a loop returns nil and WKUnknownError every other...
timothy@apple.com [Thu, 2 May 2019 23:22:40 +0000 (23:22 +0000)]
NSAttributedString conversion in a loop returns nil and WKUnknownError every other time.
https://bugs.webkit.org/show_bug.cgi?id=197523

Reviewed by Darin Adler.

Source/WebCore:

* editing/cocoa/HTMLConverter.mm:
(HTMLConverter::convert): Don't return early if m_dataSource is nil. This is already null
checked later and only needed in specific cases, it shouldn't fail the whole conversion.

Source/WebKit:

Caching the WKWebView was loading about:blank to unload the previous content.
This was causing subsequent rapid conversions to fail since the blank load
would be confused with the real content loading. Loading a blank page wasn't
really needed, it just helped keep the cached view in a cleaner state. Instead
of adding complexity to track the extra navigation, we can eliminate the blank load.
Ultimately a process swap will likely happen on the next navigation, and unused
cached views are closed quickly -- so stale content isn't held around too long.
This also avoids extra work and speeds up conversions a bit.

* UIProcess/API/Cocoa/NSAttributedString.mm:
(+[_WKAttributedStringWebViewCache cacheWebView:]): Don't load about:blank when caching.
* WebProcess/WebPage/Cocoa/WebPageCocoa.mm:
(WebKit::WebPage::getContentsAsAttributedString): Use rangeOfContents() for a fail
safe way to get the range needed.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244893 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoSetting a frame's src to a javascript URL should not run it synchronously
cdumez@apple.com [Thu, 2 May 2019 22:24:27 +0000 (22:24 +0000)]
Setting a frame's src to a javascript URL should not run it synchronously
https://bugs.webkit.org/show_bug.cgi?id=197466

Reviewed by Darin Adler.

Source/WebCore:

When an iframe's src attribute is set to a javascript URL, whether when parsing
or later on via JS, we now execute the URL's JavaScript asynchronously. We used
to execute it synchronously, which was a source of bugs and also did not match
other browsers.

I have verified that our new behavior is aligned with both Firefox and Chrome.

Note that for backward-compatibility and interoperability with Blink
(https://bugs.chromium.org/p/chromium/issues/detail?id=923585), the
"javascript:''" URL will still run synchronously. We should consider dropping
this quirk at some point.

Test: fast/dom/frame-src-javascript-url-async.html

* loader/NavigationScheduler.cpp:
(WebCore::ScheduledLocationChange::ScheduledLocationChange):
(WebCore::ScheduledLocationChange::~ScheduledLocationChange):
(WebCore::NavigationScheduler::scheduleLocationChange):
* loader/NavigationScheduler.h:
(WebCore::NavigationScheduler::scheduleLocationChange):
* loader/SubframeLoader.cpp:
(WebCore::SubframeLoader::requestFrame):

LayoutTests:

* fast/dom/frame-src-javascript-url-async-expected.txt: Added.
* fast/dom/frame-src-javascript-url-async.html: Added.
Add layout test coverage for the fact that the javascript URL is executed asynchronously
whether set during parsing or later via JS. Also makes sure that executing the javascript
URL asynchronously does not replace the frame's window. This test passes in both Chrome
and Firefox.

* imported/blink/fast/frames/navigation-in-pagehide.html:
Re-sync this test from the Blink repository.

* fast/dom/Element/id-in-frameset-expected.txt:
* fast/dom/Element/id-in-frameset.html:
* fast/dom/insertedIntoDocument-iframe-expected.txt:
* fast/dom/javascript-url-exception-isolation-expected.txt:
* fast/dom/javascript-url-exception-isolation.html:
* fast/dom/no-assert-for-malformed-js-url-attribute-expected.txt:
* fast/dom/resources/javascript-url-crash-function-iframe.html:
* fast/frames/adopt-from-created-document.html:
* fast/frames/out-of-document-iframe-has-child-frame.html:
* fast/loader/javascript-url-iframe-remove-on-navigate-async-delegate.html:
* fast/loader/javascript-url-iframe-remove-on-navigate.html:
* fast/loader/unload-mutation-crash.html:
* fast/parser/resources/set-parent-to-javascript-url.html:
* fast/parser/xml-error-adopted.xml:
* http/tests/navigation/lockedhistory-iframe-expected.txt:
* http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-javascript-url-iframe-in-iframe-expected.txt:
* http/tests/security/contentSecurityPolicy/javascript-url-allowed-expected.txt:
* http/tests/security/contentSecurityPolicy/javascript-url-blocked-by-default-src-star-expected.txt:
* http/tests/security/contentSecurityPolicy/javascript-url-blocked-expected.txt:
* http/tests/security/javascriptURL/xss-ALLOWED-from-javascript-url-sub-frame-2-level.html:
* http/tests/security/javascriptURL/xss-ALLOWED-from-javascript-url-sub-frame.html:
* http/tests/security/javascriptURL/xss-ALLOWED-to-javascript-url-from-javscript-url.html:
* imported/blink/loader/iframe-sync-loads-expected.txt:
* js/dom/call-base-resolution.html:
* platform/wk2/http/tests/security/contentSecurityPolicy/block-all-mixed-content/insecure-image-in-javascript-url-iframe-in-iframe-expected.txt:
Update / Rebaseline existing tests to reflect behavior change. I ran those tests in Firefox and Chrome to confirm that our behavior
is indeed aligned.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244892 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoSource/WebCore:
commit-queue@webkit.org [Thu, 2 May 2019 21:43:25 +0000 (21:43 +0000)]
Source/WebCore:
WebVTT: fix vertical cue alignment.
https://bugs.webkit.org/show_bug.cgi?id=136627.
<rdar://problem/49725538>

Patch by Gary Katsevman <git@gkatsev.com> on 2019-05-02
Reviewed by Eric Carlson.

Updated existing test results.

* html/track/VTTCue.cpp:
(WebCore::VTTCueBox::applyCSSProperties):

LayoutTests:
WebVTT: vertical cue text alignment is the wrong way around
https://bugs.webkit.org/show_bug.cgi?id=136627
<rdar://problem/49725538>

Patch by Gary Katsevman <git@gkatsev.com> on 2019-05-02
Reviewed by Eric Carlson.

* platform/ios/media/track/track-cue-rendering-vertical-expected.txt:
* platform/mac/TestExpectations: Unskipped test.
* platform/mac/media/track/track-cue-rendering-vertical-expected.png: Removed.
* platform/mac/media/track/track-cue-rendering-vertical-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244891 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ Sierra Release WK2 ] Layout Test fast/workers/worker-cloneport.html is flaky
tsavell@apple.com [Thu, 2 May 2019 21:20:29 +0000 (21:20 +0000)]
[ Sierra Release WK2 ] Layout Test fast/workers/worker-cloneport.html is flaky
https://bugs.webkit.org/show_bug.cgi?id=191644

Unreviewed test gardening.

* platform/mac-wk2/TestExpectations: Remove expectation.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244888 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMake both filterForRegistrableDomains() in WebKit::NetworkProcess use WebCore::Regist...
cdumez@apple.com [Thu, 2 May 2019 20:55:44 +0000 (20:55 +0000)]
Make both filterForRegistrableDomains() in WebKit::NetworkProcess use WebCore::RegistrableDomain::uncheckedCreateFromHost()
https://bugs.webkit.org/show_bug.cgi?id=197521
<rdar://problem/49651503>

Patch by John Wilander <wilander@apple.com> on 2019-05-02
Reviewed by Brent Fulgham.

Source/WebKit:

* NetworkProcess/NetworkProcess.cpp:
(WebKit::filterForRegistrableDomains):

Tools:

* TestWebKitAPI/Tests/WebCore/RegistrableDomain.cpp:
(TestWebKitAPI::TEST):
   Added a test for host strings with leading dots since they are
   often used for the domain attribute in cookies.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244884 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago-[WKWebsiteDataStore removeDataOfTypes:forDataRecords:completionHandler:] doesn't...
pvollan@apple.com [Thu, 2 May 2019 20:50:54 +0000 (20:50 +0000)]
-[WKWebsiteDataStore removeDataOfTypes:forDataRecords:completionHandler:] doesn't delete _WKWebsiteDataTypeCredentials
https://bugs.webkit.org/show_bug.cgi?id=197510
Source/WebKit:

<rdar://problem/50372338>

Reviewed by Alex Christensen.

When asked to delete persistent credentials for a set of origins, send a message to the network process with the
list of origins. The network process will then delete all persistent credentials from these origins from
NSURLCredentialStorage.

* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::removeCredentialsWithOrigins):
* NetworkProcess/NetworkProcess.h:
* NetworkProcess/NetworkProcess.messages.in:
* NetworkProcess/cocoa/NetworkProcessCocoa.mm:
(WebKit::NetworkProcess::removeCredentialsWithOrigins):
* UIProcess/WebsiteData/WebsiteDataStore.cpp:
(WebKit::WebsiteDataStore::removeData):

Tools:

Reviewed by Alex Christensen.

* TestWebKitAPI/Tests/WebKitCocoa/WKWebsiteDatastore.mm:
(TestWebKitAPI::TEST):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244883 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[CMake] Refactor WEBKIT_MAKE_FORWARDING_HEADERS into WEBKIT_COPY_FILES
don.olmstead@sony.com [Thu, 2 May 2019 20:37:41 +0000 (20:37 +0000)]
[CMake] Refactor WEBKIT_MAKE_FORWARDING_HEADERS into WEBKIT_COPY_FILES
https://bugs.webkit.org/show_bug.cgi?id=197174

Reviewed by Alex Christensen.

.:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES which is a general
purpose macro that copies files to a destination. WEBKIT_COPY_FILES removes the
requirement of a target being passed in so there is no longer any implicit dependency
between the target and the copying. Instead the dependencies need to be explicit.

This opens the macro up for use by third party libraries and for copying other files
to a destination.

* Source/cmake/WebKitMacros.cmake:

Source/JavaScriptCore:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

* CMakeLists.txt:

Source/WebCore:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

* CMakeLists.txt:

Source/WebCore/PAL:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES.

* pal/CMakeLists.txt:

Source/WebKit:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

* CMakeLists.txt:
* PlatformWin.cmake:

Source/WebKitLegacy:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES and make dependencies
for framework headers explicit.

* PlatformWin.cmake:

Source/WTF:

Replace WEBKIT_MAKE_FORWARDING_HEADERS with WEBKIT_COPY_FILES.

* wtf/CMakeLists.txt:

Tools:

Make TestJSC dependent on JavaScriptCorePrivateFrameworkHeaders.

* TestWebKitAPI/PlatformGTK.cmake:
* TestWebKitAPI/PlatformWPE.cmake:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244881 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[WebAuthN] Add a quirk for google.com when processing AppID extension
jiewen_tan@apple.com [Thu, 2 May 2019 19:15:08 +0000 (19:15 +0000)]
[WebAuthN] Add a quirk for google.com when processing AppID extension
https://bugs.webkit.org/show_bug.cgi?id=196046
<rdar://problem/49088479>

Reviewed by Brent Fulgham.

Relaxing the same site restriction on AppID while in google.com and any
of its subdomains to allow two www.gstatic.com AppIDs to slip in.

Covered by manual tests on Google.com.

* Modules/webauthn/AuthenticatorCoordinator.cpp:
(WebCore::AuthenticatorCoordinatorInternal::needsAppIdQuirks):
(WebCore::AuthenticatorCoordinatorInternal::processAppIdExtension):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244879 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[macOS] Fix syscall sandbox violation
pvollan@apple.com [Thu, 2 May 2019 18:09:44 +0000 (18:09 +0000)]
[macOS] Fix syscall sandbox violation
https://bugs.webkit.org/show_bug.cgi?id=197517
<rdar://problem/49945031>

Reviewed by Brent Fulgham.

A syscall needs to be added to the sandbox on macOS.

* WebProcess/com.apple.WebProcess.sb.in:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244877 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed fix for non-unified build after r244853.
ross.kirsling@sony.com [Thu, 2 May 2019 17:59:02 +0000 (17:59 +0000)]
Unreviewed fix for non-unified build after r244853.

* page/SecurityOrigin.cpp:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244876 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoSafe browsing warning should update colors when a user switches between light and...
achristensen@apple.com [Thu, 2 May 2019 17:38:07 +0000 (17:38 +0000)]
Safe browsing warning should update colors when a user switches between light and dark appearance
https://bugs.webkit.org/show_bug.cgi?id=197443
<rdar://problem/49883917>

Reviewed by Tim Horton.

We draw the WKSafeBrowsingExclamationPoint ourselves, so we need to call setNeedsDisplay when viewDidChangeEffectiveAppearance is called.
Instead of setting NSView.layer.backgroundColor we need to make an NSView subclass I call WKSafeBrowsingBox and we need to set its layer's
backgroundColor in updateLayer, otherwise the CGColor isn't updated from the NSColor.

* UIProcess/Cocoa/WKSafeBrowsingWarning.h:
* UIProcess/Cocoa/WKSafeBrowsingWarning.mm:
(colorForItem):
(-[WKSafeBrowsingExclamationPoint viewDidChangeEffectiveAppearance]):
(-[WKSafeBrowsingBox setSafeBrowsingBackgroundColor:]):
(-[WKSafeBrowsingBox updateLayer]):
(-[WKSafeBrowsingWarning initWithFrame:safeBrowsingWarning:completionHandler:]):
(-[WKSafeBrowsingWarning addContent]):
(-[WKSafeBrowsingWarning showDetailsClicked]):
(setBackground): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244875 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoBuild fix for internal branch.
achristensen@apple.com [Thu, 2 May 2019 17:35:23 +0000 (17:35 +0000)]
Build fix for internal branch.

* TestWebKitAPI/Tests/WebKitCocoa/Challenge.mm:
RetainPtr.h was not explicitly included.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244874 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agowebkitpy: Ignore errors when shutting down an already shutdown simulator
jbedard@apple.com [Thu, 2 May 2019 17:01:35 +0000 (17:01 +0000)]
webkitpy: Ignore errors when shutting down an already shutdown simulator
<https://bugs.webkit.org/show_bug.cgi?id=197514>
<rdar://problem/50390247>

Reviewed by Lucas Forschler.

* Scripts/webkitpy/xcode/simulated_device.py:
(SimulatedDevice._shut_down):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244873 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed rollout of r244862.
msaboff@apple.com [Thu, 2 May 2019 16:17:53 +0000 (16:17 +0000)]
Unreviewed rollout of r244862.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244872 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed GTK+ gardening after r244868.
magomez@igalia.com [Thu, 2 May 2019 15:21:58 +0000 (15:21 +0000)]
Unreviewed GTK+ gardening after r244868.

* platform/gtk/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244871 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[Win10] Some tests are failing only on specific machines
pvollan@apple.com [Thu, 2 May 2019 14:57:20 +0000 (14:57 +0000)]
[Win10] Some tests are failing only on specific machines
https://bugs.webkit.org/show_bug.cgi?id=197509

Unreviewed test gardening.

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244870 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[GTK][WPE] Disable "thin", "thick", "medium" values of mfrac@linethickness at runtime
commit-queue@webkit.org [Thu, 2 May 2019 10:12:14 +0000 (10:12 +0000)]
[GTK][WPE] Disable "thin", "thick", "medium" values of mfrac@linethickness at runtime
https://bugs.webkit.org/show_bug.cgi?id=196142

This patch introduces some experimental runtime flag to let users
disable MathML features that are removed from MathML Core [1]. For now,
LayoutTests/imported/w3c:

these features are only disabled on GTK and WPE ports. This patch also
adds a condition to disable "thin", "thick", "medium" values of
mfrac@linethickness at runtime as agreed in [2].

[1] https://mathml-refresh.github.io/mathml-core/
[2] https://github.com/mathml-refresh/mathml/issues/4

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

* web-platform-tests/mathml/presentation-markup/fractions/frac-linethickness-001.html:
Explicitly enable deprecated features in this test. It should be moved
back to WebKit's test suite if/when it is updated in the upstream WPT
repository so also add a reminder for the next synchronization.

Source/WebCore:

these features are only disabled on GTK and WPE ports. This patch also
adds a condition to disable "thin", "thick", "medium" values of
mfrac@linethickness at runtime as agreed in [2].

[1] https://mathml-refresh.github.io/mathml-core/
[2] https://github.com/mathml-refresh/mathml/issues/4

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

No new tests, covered by frac-linethickness-0001.html

* mathml/MathMLFractionElement.cpp:
(WebCore::MathMLFractionElement::lineThickness): Just do standard
parsing for MathML lengths when non-core MathML features are disabled.
* page/Settings.yaml: Add WebCore setting.

Source/WebKit:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

* Shared/WebPreferences.yaml:
* Shared/WebPreferencesDefaultValues.h:

Source/WebKitLegacy/mac:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

* WebView/WebPreferenceKeysPrivate.h:
* WebView/WebPreferences.mm:
(+[WebPreferences initialize]):
(-[WebPreferences coreMathMLEnabled]):
(-[WebPreferences setCoreMathMLEnabled:]):
* WebView/WebPreferencesPrivate.h:
* WebView/WebView.mm:
(-[WebView _preferencesChanged:]):

Source/WebKitLegacy/win:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

* Interfaces/IWebPreferencesPrivate.idl:
* WebPreferenceKeysPrivate.h:
* WebPreferences.cpp:
(WebPreferences::initializeDefaultSettings):
(WebPreferences::coreMathMLEnabled):
(WebPreferences::setCoreMathMLEnabled):
* WebPreferences.h:
* WebView.cpp:
(WebView::notifyPreferencesChanged):

Tools:

these features are only disabled on GTK and WPE ports.

[1] https://mathml-refresh.github.io/mathml-core/

Patch by Frederic Wang <fwang@igalia.com> on 2019-05-02
Reviewed by Rob Buis.

* DumpRenderTree/TestOptions.cpp:
(TestOptions::TestOptions):
* DumpRenderTree/TestOptions.h:
* DumpRenderTree/mac/DumpRenderTree.mm:
(setWebPreferencesForTestOptions):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244869 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, fix test failures after r242992.
drousso@apple.com [Thu, 2 May 2019 06:27:07 +0000 (06:27 +0000)]
Unreviewed, fix test failures after r242992.
<rdar://problem/50383600>

* inspector/console/console-screenshot.html:
* inspector/console/console-screenshot-expected.txt:
Don't print the dataURL in the expectation file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244868 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd the ability to import WebKitAdditions snippets in WebKit.apinotes
wenson_hsieh@apple.com [Thu, 2 May 2019 04:20:38 +0000 (04:20 +0000)]
Add the ability to import WebKitAdditions snippets in WebKit.apinotes
https://bugs.webkit.org/show_bug.cgi?id=197487
<rdar://problem/50389085>

Reviewed by Tim Horton.

* Shared/API/Cocoa/WebKit.apinotes: Import a new WebKitAdditions file.

* WebKit.xcodeproj/project.pbxproj: Check all files in Headers/ and PrivateHeaders/ that need replacement, not
simply anything ending with ".h".

* mac/replace-webkit-additions-includes.py:
(main): Teach the replacement script to handle files in Headers/ or PrivateHeaders/ that don't end in ".h".

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244866 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoBaseline JIT should do argument value profiling after checking for stack overflow
sbarati@apple.com [Thu, 2 May 2019 03:10:43 +0000 (03:10 +0000)]
Baseline JIT should do argument value profiling after checking for stack overflow
https://bugs.webkit.org/show_bug.cgi?id=197052
<rdar://problem/50009602>

Reviewed by Yusuke Suzuki.

JSTests:

* stress/check-stack-overflow-before-value-profiling-arguments.js: Added.

Source/JavaScriptCore:

Otherwise, we may do value profiling without running a write barrier, which
is against the rules of how we do value profiling.

* jit/JIT.cpp:
(JSC::JIT::compileWithoutLinking):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244865 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[JSC] Inlining Getter/Setter should care availability of ad-hocly constructed frame
ysuzuki@apple.com [Thu, 2 May 2019 02:40:44 +0000 (02:40 +0000)]
[JSC] Inlining Getter/Setter should care availability of ad-hocly constructed frame
https://bugs.webkit.org/show_bug.cgi?id=197405

Reviewed by Saam Barati.

JSTests:

* stress/getter-setter-inlining-should-emit-movhint.js: Added.
(foo):
(test):
(i.o.get f):
(i.o.set f):

Source/JavaScriptCore:

When inlining getter and setter calls, we setup a stack frame which does not appear in the bytecode.
Because Inlining can switch on executable, we could have a graph like this.

BB#0
    ...
    30: GetSetter
    31: MovHint(loc10)
    32: SetLocal(loc10)
    33: MovHint(loc9)
    34: SetLocal(loc9)
    ...
    37: GetExecutable(@30)
    ...
    41: Switch(@37)

BB#2
    42: GetLocal(loc12, bc#7 of caller)
    ...
    --> callee: loc9 and loc10 are arguments of callee.
      ...
      <HERE, exit to callee, loc9 and loc10 are required in the bytecode>

When we prune OSR availability at the beginning of BB#2 (bc#7 in the caller), we prune loc9 and loc10's liveness because the caller does not actually have loc9 and loc10.
However, when we begin executing the callee, we need OSR exit to be aware of where it can recover the arguments to the setter, loc9 and loc10.

This patch inserts MovHint at the beginning of callee for a getter / setter stack frame to make arguments (loc9 and loc10 in the above example) recoverable from OSR exit.
We also move arity fixup DFG nodes from the caller to the callee, since moved arguments are not live in the caller too.

Interestingly, this fix also reveals the existing issue in LiveCatchVariablePreservationPhase. We emitted Flush for |this| of InlineCallFrame blindly if we saw InlineCallFrame
inside a block which is covered by catch handler. But this is wrong because inlined function can finish its execution within the block, and |this| is completely unrelated to
the catch handler if the catch handler is in the outer callee. We already collect all the live locals at the catch handler. And this locals must include arguments too if the
catch handler is in inlined function. So, we should not emit Flush for each |this| of seen InlineCallFrame. This emitted Flush may connect unrelated locals in the catch handler
to the locals that is only defined and used in the inlined function, and it leads to the results like DFG says the local is live while the bytecode says the local is dead.
This results in reading and using garbage in OSR entry because DFG OSR entry needs to fill live DFG values from the stack.

* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::inlineCall):
(JSC::DFG::ByteCodeParser::handleGetById):
(JSC::DFG::ByteCodeParser::handlePutById):
* dfg/DFGLiveCatchVariablePreservationPhase.cpp:
(JSC::DFG::LiveCatchVariablePreservationPhase::handleBlockForTryCatch):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244864 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[WebAuthN] Adopt SecurityOrigin::isMatchingRegistrableDomainSuffix()
jiewen_tan@apple.com [Thu, 2 May 2019 02:20:51 +0000 (02:20 +0000)]
[WebAuthN] Adopt SecurityOrigin::isMatchingRegistrableDomainSuffix()
https://bugs.webkit.org/show_bug.cgi?id=197481

Reviewed by Brent Fulgham.

Source/WebCore:

This patch implements Step 6-7 from:
https://www.w3.org/TR/webauthn/#createCredential,
https://www.w3.org/TR/webauthn/#discover-from-external-source.

Test: http/wpt/webauthn/public-key-credential-ip-address.html

* Modules/webauthn/AuthenticatorCoordinator.cpp:
(WebCore::AuthenticatorCoordinator::create const):
(WebCore::AuthenticatorCoordinator::discoverFromExternalSource const):

LayoutTests:

* http/wpt/webauthn/public-key-credential-create-failure.https.html:
* http/wpt/webauthn/public-key-credential-get-failure.https.html:
* http/wpt/webauthn/public-key-credential-ip-address-expected.txt: Added.
* http/wpt/webauthn/public-key-credential-ip-address.html: Added.
* http/wpt/webauthn/resources/public-key-credential-ip-address.https.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244863 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoASSERTION FAILED: !m_needExceptionCheck with --validateExceptionChecks=1; ProxyObject...
msaboff@apple.com [Thu, 2 May 2019 02:04:07 +0000 (02:04 +0000)]
ASSERTION FAILED: !m_needExceptionCheck with --validateExceptionChecks=1; ProxyObject.getOwnPropertySlotCommon/JSFunction.callerGetter
https://bugs.webkit.org/show_bug.cgi?id=197485

Reviewed by Saam Barati.

JSTests:

New test.

* stress/proxy-getOwnPropertySlots-exceptionChecks.js: Added.
(foo):

Source/JavaScriptCore:

Added an EXCEPTION_ASSERT after call to getOwnPropertySlot().

* runtime/JSObject.cpp:
(JSC::JSObject::getOwnPropertyDescriptor):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244862 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRemoteInspector::updateAutomaticInspectionCandidate should have a default implementation.
ross.kirsling@sony.com [Thu, 2 May 2019 00:25:46 +0000 (00:25 +0000)]
RemoteInspector::updateAutomaticInspectionCandidate should have a default implementation.
https://bugs.webkit.org/show_bug.cgi?id=197439

Reviewed by Devin Rousso.

On non-Cocoa platforms, automatic inspection is not currently implemented,
so updateAutomaticInspectionCandidate falls back to the logic of updateTarget.
This logic already existed in three places, so refactor it into a common private method
and allow our websocket-based RWI implementation to make use of it too.

* inspector/remote/RemoteInspector.cpp:
(Inspector::RemoteInspector::updateTarget):
(Inspector::RemoteInspector::updateTargetMap):
(Inspector::RemoteInspector::updateAutomaticInspectionCandidate):
* inspector/remote/RemoteInspector.h:
* inspector/remote/cocoa/RemoteInspectorCocoa.mm:
(Inspector::RemoteInspector::updateAutomaticInspectionCandidate):
* inspector/remote/glib/RemoteInspectorGlib.cpp:
(Inspector::RemoteInspector::updateAutomaticInspectionCandidate): Deleted.
* inspector/remote/socket/RemoteInspectorSocket.cpp:
(Inspector::RemoteInspector::updateAutomaticInspectionCandidate): Deleted.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244861 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd back hasNullReferences() assert in Document::updateIsPlayingMedia
youenn@apple.com [Thu, 2 May 2019 00:20:38 +0000 (00:20 +0000)]
Add back hasNullReferences() assert in Document::updateIsPlayingMedia
https://bugs.webkit.org/show_bug.cgi?id=197477

Reviewed by Eric Carlson.

AudioContext had a special scriptExecutionContext() getter
that was returning nullptr when AudioContext is stopped.
Instead, make this getter behave as all other scriptExecutionContext() getters
and make sure existing call sites of the getter are fine with the change.

Covered by existing tests.

* Modules/webaudio/AudioContext.cpp:
(WebCore::AudioContext::AudioContext):
No need to check whether offline since this constructor is for non offline AudioContext.
(WebCore::AudioContext::scriptExecutionContext const):
(WebCore::AudioContext::fireCompletionEvent):
* dom/Document.cpp:
(WebCore::Document::updateIsPlayingMedia):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244860 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[PlayStation] Update port for WEBKIT_EXECUTABLE changes
stephan.szabo@sony.com [Thu, 2 May 2019 00:09:50 +0000 (00:09 +0000)]
[PlayStation] Update port for WEBKIT_EXECUTABLE changes
https://bugs.webkit.org/show_bug.cgi?id=197483

Reviewed by Don Olmstead.

* Source/cmake/OptionsPlayStation.cmake:
Update wrapping for configuring executables' playstation
specific data to new WEBKIT_EXECUTABLE macro.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244859 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, fix test failures after r242809.
drousso@apple.com [Thu, 2 May 2019 00:00:45 +0000 (00:00 +0000)]
Unreviewed, fix test failures after r242809.
<rdar://problem/50383600>

* inspector/canvas/recording-html-2d.html:
* inspector/canvas/recording-html-2d-expected.txt:
Don't print the dataURL in the expectation file.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244858 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoTestWebKitAPI config.h should be aware of what suite is being built
don.olmstead@sony.com [Wed, 1 May 2019 23:47:28 +0000 (23:47 +0000)]
TestWebKitAPI config.h should be aware of what suite is being built
https://bugs.webkit.org/show_bug.cgi?id=196583

Reviewed by Konstantin Tokarev.

Migrate to WEBKIT_EXECUTABLE to define the Test* executables. As an added
benefit the macro defines BUILDING_* for the executable which allows the includes
of the config.h to be tailored to the test suite being built. Because of that the
dependencies of the Test* executable are specific to the test suite which may
speed up the build.

Some source files were including "PlatformUtilities.h" which includes WebKit headers
when "Utilities.h" is what was needed.

* TestWebKitAPI/CMakeLists.txt:
* TestWebKitAPI/PlatformGTK.cmake:
* TestWebKitAPI/PlatformJSCOnly.cmake:
* TestWebKitAPI/PlatformMac.cmake:
* TestWebKitAPI/PlatformPlayStation.cmake:
* TestWebKitAPI/PlatformUtilities.h:
* TestWebKitAPI/PlatformWPE.cmake:
* TestWebKitAPI/PlatformWin.cmake:
* TestWebKitAPI/Tests/WTF/RefPtr.cpp:
* TestWebKitAPI/Tests/WebCore/ContentExtensions.cpp:
* TestWebKitAPI/Tests/WebCore/FileMonitor.cpp:
* TestWebKitAPI/Tests/WebCore/LineBreaking.mm:
* TestWebKitAPI/Tests/WebCore/cocoa/SharedBuffer.mm:
* TestWebKitAPI/Tests/WebCore/cocoa/WebCoreNSURLSession.mm:
* TestWebKitAPI/config.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244857 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUpdate WebGPU class names based on sketch.idl
justin_fan@apple.com [Wed, 1 May 2019 23:21:59 +0000 (23:21 +0000)]
Update WebGPU class names based on sketch.idl
https://bugs.webkit.org/show_bug.cgi?id=194260

Reviewed by Dean Jackson.

Update all exposed Web GPU interface names to GPU* prefix.

Source/WebCore:

Existing Web GPU tests updated to expect new names.

* Modules/webgpu/WebGPU.idl:
* Modules/webgpu/WebGPUAdapter.idl:
* Modules/webgpu/WebGPUBindGroup.idl:
* Modules/webgpu/WebGPUBindGroupLayout.idl:
* Modules/webgpu/WebGPUBuffer.idl:
* Modules/webgpu/WebGPUDevice.idl:
* Modules/webgpu/WebGPUInputStepMode.h: Removed.
* Modules/webgpu/WebGPUPipelineLayout.idl:
* Modules/webgpu/WebGPUProgrammablePassEncoder.idl:
* Modules/webgpu/WebGPUQueue.idl:
* Modules/webgpu/WebGPURenderPassEncoder.idl:
* Modules/webgpu/WebGPURenderPipeline.idl:
* Modules/webgpu/WebGPUSampler.idl:
* Modules/webgpu/WebGPUTexture.idl:
* Modules/webgpu/WebGPUTextureView.idl:
* WebCore.xcodeproj/project.pbxproj:
* bindings/js/WebCoreBuiltinNames.h:

LayoutTests:

* webgpu/adapter-options.html:
* webgpu/bind-groups-expected.txt:
* webgpu/bind-groups.html:
* webgpu/buffer-resource-triangles.html:
* webgpu/js/webgpu-functions.js:
(beginBasicRenderPass):
* webgpu/map-read-buffers.html:
* webgpu/map-write-buffers.html:
* webgpu/pipeline-layouts-expected.txt:
* webgpu/pipeline-layouts.html:
* webgpu/queue-creation-expected.txt:
* webgpu/queue-creation.html:
* webgpu/render-command-encoding-expected.txt:
* webgpu/render-command-encoding.html:
* webgpu/render-pipelines-expected.txt:
* webgpu/render-pipelines.html:
* webgpu/textures-textureviews.html:
* webgpu/webgpu-enabled.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244856 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] Unreviewed test gardening.
zalan@apple.com [Wed, 1 May 2019 23:07:35 +0000 (23:07 +0000)]
[iOS] Unreviewed test gardening.

* platform/ios-wk2/fast/block/margin-collapse/103-expected.txt:
* platform/ios/compositing/overflow/theme-affects-visual-overflow-expected.txt:
* platform/ios/fast/block/margin-collapse/103-expected.txt:
* platform/ios/fast/parser/bad-xml-slash-expected.txt:
* platform/ios/tables/mozilla/bugs/bug4527-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244855 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoREGRESSION(r244494): editing/pasteboard/5761530-1.html is flaky due to a simple line...
wenson_hsieh@apple.com [Wed, 1 May 2019 22:34:03 +0000 (22:34 +0000)]
REGRESSION(r244494): editing/pasteboard/5761530-1.html is flaky due to a simple line layout bug
https://bugs.webkit.org/show_bug.cgi?id=197438
<rdar://problem/50348091>

Reviewed by Zalan Bujtas.

Work around the simple line layout bug in the flaky layout test by disabling simple line layout.

* editing/pasteboard/5761530-1.html:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244854 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMove Document::domainIsRegisterable to SecurityOrigin::isMatchingRegistrableDomainSuffix
jiewen_tan@apple.com [Wed, 1 May 2019 22:08:00 +0000 (22:08 +0000)]
Move Document::domainIsRegisterable to SecurityOrigin::isMatchingRegistrableDomainSuffix
https://bugs.webkit.org/show_bug.cgi?id=181950
<rdar://problem/43357371>

Reviewed by Brent Fulgham.

Source/WebCore:

This patch moves Document::domainIsRegisterable to SecurityOrigin::isMatchingRegistrableDomainSuffix
to be more aligned with the HTML standard:
https://html.spec.whatwg.org/multipage/origin.html#is-a-registrable-domain-suffix-of-or-is-equal-to.
Besides that, it also removes redundant codes within the original method that is also done in
OriginAccessEntry::matchesOrigin.

Covered by new API tests.

* dom/Document.cpp:
(WebCore::Document::setDomain):
(WebCore::Document::domainIsRegisterable const): Deleted.
* dom/Document.h:
* page/SecurityOrigin.cpp:
(WebCore::SecurityOrigin::isMatchingRegistrableDomainSuffix const):
* page/SecurityOrigin.h:

Tools:

* TestWebKitAPI/Tests/WebCore/SecurityOrigin.cpp:
(TestWebKitAPI::TEST_F):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244853 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRemove webkitpy and bindings EWS queues from dashboard
aakash_jain@apple.com [Wed, 1 May 2019 21:40:25 +0000 (21:40 +0000)]
Remove webkitpy and bindings EWS queues from dashboard
https://bugs.webkit.org/show_bug.cgi?id=197475

Reviewed by Alexey Proskuryakov.

* BuildSlaveSupport/build.webkit.org-config/public_html/dashboard/Scripts/BubbleQueueServer.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244852 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] Element::focus and Element::scrollIntoView do not clamp scroll positions
rniwa@webkit.org [Wed, 1 May 2019 21:38:01 +0000 (21:38 +0000)]
[iOS] Element::focus and Element::scrollIntoView do not clamp scroll positions
https://bugs.webkit.org/show_bug.cgi?id=197211

Reviewed by Simon Fraser.

Source/WebCore:

Fixed the bug that Element::focus and Element::scrollIntoView were not clamping scroll offsets,
which causes scrollTop etc... to return a bogus negative scrolling offset.

Unfortunately, we can't just use FrameView's ScrollableArea::constrainScrollPosition since
scrollRectToVisible relies on the visible rect being expanded by the content insets in order to scroll to
a position within the content insets of UIScrollView; e.g. revealing the top of the page as the center.
We manually expand minimumScrollPosition() and maximumScrollPosition() by the content insets instead.

Tests: fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top.html
       fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top.html
       fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top.html

* rendering/RenderLayer.cpp:
(WebCore::RenderLayer::scrollRectToVisible):

LayoutTests:

Added tests for scrolling via Element.prototype.focus and Element.prototype.scrollIntoView.

* fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top-expected.txt: Added.
* fast/scrolling/ios/programmatic-scroll-via-focus-should-clamp-top.html: Added.
* fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top-expected.txt: Added.
* fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-inside-iframe-should-clamp-top.html: Added.
* fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top-expected.txt: Added.
* fast/scrolling/ios/programmatic-scroll-via-scrollIntoView-should-clamp-top.html: Added.
* platform/ios-wk2/fast/dom/focus-contenteditable-expected.txt: Rebaselined the output now that the scroll position
is properly clamped.
* platform/ios-wk2/imported/w3c/web-platform-tests/css/cssom-view/elementFromPoint-expected.txt: Rebaselined as one
more test case is passing.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244851 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoProtect against null crash in fetchDiskCacheEntries
achristensen@apple.com [Wed, 1 May 2019 21:19:24 +0000 (21:19 +0000)]
Protect against null crash in fetchDiskCacheEntries
https://bugs.webkit.org/show_bug.cgi?id=197399
<rdar://problem/47759337>

Reviewed by Antti Koivisto.

If the call to Entry::decodeStorageRecord returns nullptr in Storage::traverse, do not call the traverseHandler, which expects
that when it is called with nullptr that is the end of the traversal.

* NetworkProcess/cache/NetworkCache.cpp:
(WebKit::NetworkCache::Cache::traverse):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244850 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] Add a version of viewport shrink-to-fit heuristics that preserves page layout
wenson_hsieh@apple.com [Wed, 1 May 2019 21:08:38 +0000 (21:08 +0000)]
[iOS] Add a version of viewport shrink-to-fit heuristics that preserves page layout
https://bugs.webkit.org/show_bug.cgi?id=197342
<rdar://problem/50063091>

Reviewed by Tim Horton.

Source/WebCore:

Adds support for a new shrink-to-fit heuristic that attempts to lay out the contents of the page at a larger
width in order to shrink content to fit the viewport. See WebKit ChangeLog for more details.

Tests: fast/viewport/ios/shrink-to-fit-content-constant-width.html
       fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint.html
       fast/viewport/ios/shrink-to-fit-content-no-viewport.html
       fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow.html
       fast/viewport/ios/shrink-to-fit-content-temporary-overflow.html

* page/ViewportConfiguration.cpp:
(WebCore::ViewportConfiguration::setMinimumEffectiveDeviceWidth):
(WebCore::ViewportConfiguration::setIsKnownToLayOutWiderThanViewport):
(WebCore::ViewportConfiguration::description const):
* page/ViewportConfiguration.h:
(WebCore::ViewportConfiguration::canIgnoreScalingConstraints const):
(WebCore::ViewportConfiguration::minimumEffectiveDeviceWidth const):

Add several new getters and setters in ViewportConfiguration.

(WebCore::ViewportConfiguration::isKnownToLayOutWiderThanViewport const):
(WebCore::ViewportConfiguration::shouldIgnoreMinimumEffectiveDeviceWidth const):

Importantly, only allow ignoring the minimum effective device width in webpages with responsive viewports, if
they also have *not* laid out wider than the viewport.

(WebCore::ViewportConfiguration::setForceAlwaysUserScalable):

Source/WebKit:

This patch introduces a new shrink-to-fit heuristic that attempts to lay out the contents of the page at a
larger width in order to shrink content to fit the viewport. This is similar to existing shrink-to-fit behaviors
used for viewport sizing in multitasking mode, except that it not only scales the view, but additionally expands
the layout size, such that the overall layout of the page is preserved. In fact, the reason we ended up
reverting the existing flavor of shrink-to-fit in all cases except for multitasking was that page layout was not
preserved, which caused elements that poke out of the viewport to make the rest of the page look out of
proportion — see <rdar://problem/23818102> and related radars.

Covered by 5 new layout tests, and by adjusting a couple of existing layout tests. See comments below for more
details.

* Platform/Logging.h:

Add a new ViewportSizing logging channel. This will only log on pages that overflow the viewport and shrink to
fit as a result.

* Shared/WebPreferences.yaml:

Turn IgnoreViewportScalingConstraints off by default. This preference currently controls whether we allow
shrink-to-fit behaviors, and is only used by Safari when it is in multitasking mode. The value of this
preference is currenly *on* by default, and is turned off almost immediately during every page load after the
first visible content rect update, wherein visibleContentRectUpdateInfo.allowShrinkToFit() is false.

However, this sometimes causes a brief jitter during page load; to fix this, make the default value for
IgnoreViewportScalingConstraints false, and change the logic in WebPage::updateVisibleContentRects to
setCanIgnoreScalingConstraints to true if either the IgnoreViewportScalingConstraints preference (not only
affected by an internal debug switch) is true, or WKWebView SPI is used to enable the behavior.

* WebProcess/WebCoreSupport/WebFrameLoaderClient.cpp:
(WebKit::WebFrameLoaderClient::dispatchDidFinishDocumentLoad):
(WebKit::WebFrameLoaderClient::dispatchDidFinishLoad):

Add a new hook for WebFrameLoaderClient to call into WebPage when document load finishes. Also, tweak
dispatchDidFinishLoad to take a WebFrame& instead of a WebFrame* in a drive-by fix (the frame is assumed to be
non-null anyways).

* WebProcess/WebPage/WebPage.cpp:
(WebKit::WebPage::didCommitLoad):
(WebKit::WebPage::didFinishDocumentLoad):
(WebKit::WebPage::didFinishLoad):

When finishing document load or finishing the overall load, kick off the shrink-to-fit timer; when committing a
load, cancel the timer.

* WebProcess/WebPage/WebPage.h:
* WebProcess/WebPage/ios/WebPageIOS.mm:
(WebKit::WebPage::setViewportConfigurationViewLayoutSize):

Don't allow the minimum effective device width from the client to stomp over any minimum effective device width
set as a result of the new shrink-to-fit heuristic; on some pages that load quickly, this can result in a race
where the minimum effective device width (i.e. a value that lower-bounds the minimum layout width) is first set
by the shrink-to-fit heuristic, and then set to an incorrect value by the client.

In the near future, web view SPI used to set the minimum effective device width should actually be removed
altogether, since the new shrink-to-fit heuristic supersedes any need for the client to fiddle with the minimum
effective device width.

(WebKit::WebPage::dynamicViewportSizeUpdate):

When performing a dynamic viewport size update, additionally re-run the shrink-to-fit heuristic. This allows
the minimum layout size of the viewport to be updated, if necessary. An example of where this matters is when a
web page is *below* a tablet/desktop layout breakpoint in portrait device orientation, but then exceeds this
layout breakpoint in landscape orientation. In this scenario, rotating the device should swap between these two
page layouts.

(WebKit::WebPage::resetViewportDefaultConfiguration):
(WebKit::WebPage::scheduleShrinkToFitContent):
(WebKit::WebPage::shrinkToFitContentTimerFired):
(WebKit::WebPage::immediatelyShrinkToFitContent):

Leverage the existing capability for a viewport to have a "minimum effective device width" to grant the viewport
a larger layout size than it would normally have, and then scale down to fit within the bounds of the view. One
challenge with this overall approach is that laying out at a larger width may cause the page to lay out even
wider in response, which may actually worsen horizontal scrolling. To mitigate this, we only attempt to lay out
at the current content width once; if laying out at this width reduced the amount of horizontal scrolling by any
amount, then proceed with this layout width; otherwise, revert to the previous layout width.

(WebKit::WebPage::shouldIgnoreMetaViewport const):

Pull some common logic out into a readonly getter.

(WebKit::WebPage::updateVisibleContentRects):

See the comment below WebPreferences.yaml, above.

LayoutTests:

Introduces new layout tests, and adjusts some existing tests. See comments below.

* fast/viewport/ios/shrink-to-fit-content-constant-width-expected.txt: Added.
* fast/viewport/ios/shrink-to-fit-content-constant-width.html: Added.

Add a new layout test to exercise the scenario where a constant width viewport narrower than the view is used.

* fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint-expected.txt: Added.
* fast/viewport/ios/shrink-to-fit-content-large-width-breakpoint.html: Added.

Add a new layout test to exercise the scenario where a responsive website that lays out larger than the view
width ends up with even more horizontal scrolling when laying out at the initial content width. In this
scenario, we shouldn't try to expand the viewport to try and encompass the content width, since that would only
induce even worse horizontal scrolling.

* fast/viewport/ios/shrink-to-fit-content-no-viewport-expected.txt: Added.
* fast/viewport/ios/shrink-to-fit-content-no-viewport.html: Added.

Add a new layout test for the case where there is no viewport, but content lays out wider than the view.

* fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow-expected.txt: Added.
* fast/viewport/ios/shrink-to-fit-content-responsive-viewport-with-horizontal-overflow.html: Added.

Add a new layout test for the case where the page has opted for a responsive viewport (device-width, initial
scale 1), but has laid out wider than the viewport anyways. In this case, we want to shrink the contents down to
fit inside the view.

* fast/viewport/ios/shrink-to-fit-content-temporary-overflow-expected.txt: Added.
* fast/viewport/ios/shrink-to-fit-content-temporary-overflow.html: Added.

Add a new layout test to exercise the case where, during page load, content width temporarily increases, and
then decreases such that it once again fits within the viewport. In this case, we don't want to expand the
viewport to be as wide as the large temporary width of the page.

* fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden-expected.txt:
* fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden.html:
* fast/viewport/ios/width-is-device-width-overflowing-expected.txt:
* fast/viewport/ios/width-is-device-width-overflowing.html:

Tweak these 2 existing layout tests to include "shrink-to-fit=no", to prevent the new heuristics from shrinking
the page to fit on device classes that use native viewports by default.

* platform/ipad/fast/viewport/ios/width-is-device-width-overflowing-body-overflow-hidden-expected.txt:
* platform/ipad/fast/viewport/ios/width-is-device-width-overflowing-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244849 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] Star rating is covered with a black circle when writing a review on Yelp
zalan@apple.com [Wed, 1 May 2019 20:46:20 +0000 (20:46 +0000)]
[iOS] Star rating is covered with a black circle when writing a review on Yelp
https://bugs.webkit.org/show_bug.cgi?id=197469
<rdar://problem/48094446>

Reviewed by Dean Jackson.

Source/WebCore:

This patch moves the background painting of the radio/checkbox form controls in checked state to RenderTheme.
It enables content authors to disable default appearance using -webkit-appearance: none (it is also inline with what we do on macOS).

Test: fast/forms/radio-and-checkbox-checked-with-no-appearance.html

* css/html.css:
(input:matches([type="checkbox"], [type="radio"]):checked):
* rendering/RenderThemeIOS.mm:
(WebCore::RenderThemeIOS::paintCheckboxDecorations):
(WebCore::RenderThemeIOS::paintRadioDecorations):

LayoutTests:

* fast/forms/radio-and-checkbox-checked-with-no-appearance-expected.html: Added.
* fast/forms/radio-and-checkbox-checked-with-no-appearance.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244848 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoREGRESSION: Layout test imported/w3c/web-platform-tests/resource-timing/resource...
ryanhaddad@apple.com [Wed, 1 May 2019 20:35:35 +0000 (20:35 +0000)]
REGRESSION: Layout test imported/w3c/web-platform-tests/resource-timing/resource-timing-level1.sub.html is frequently failing on Mac-ews
https://bugs.webkit.org/show_bug.cgi?id=197473

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244844 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRemove webkitpy and bindings test queues from old EWS
aakash_jain@apple.com [Wed, 1 May 2019 20:10:49 +0000 (20:10 +0000)]
Remove webkitpy and bindings test queues from old EWS
https://bugs.webkit.org/show_bug.cgi?id=197468

Reviewed by Alexey Proskuryakov.

* QueueStatusServer/config/queues.py:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244840 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWin10 failing or flakey tests need to be given expectations to allow for new EWS
tsavell@apple.com [Wed, 1 May 2019 19:40:05 +0000 (19:40 +0000)]
Win10 failing or flakey tests need to be given expectations to allow for new EWS
https://bugs.webkit.org/show_bug.cgi?id=197471

Unreviewed test Gardening

* platform/win/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244839 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoREGRESSION (r244182): RenderingUpdate should not be scheduled for invisible pages
commit-queue@webkit.org [Wed, 1 May 2019 19:31:06 +0000 (19:31 +0000)]
REGRESSION (r244182): RenderingUpdate should not be scheduled for invisible pages
https://bugs.webkit.org/show_bug.cgi?id=197451

Patch by Said Abou-Hallawa <sabouhallawa@apple.com> on 2019-05-01
Reviewed by Simon Fraser.

Before r244182, some web pages never need to schedule a RenderingUpdate.
Only pages with rAF callbacks, web animations, intersection and resize
observers needed to do so. After r244182, all pages have to schedule a
RenderingUpdate when a page rendering update is required.

When Safari opens, it create a 'blank' web page. The blank page will not
be visible unless the user selects to show the 'Empty page' in the new
tab. Although the blank page is not visible, the loader needs to resolveStyle()
which requires to scheduleLayerFlushNow().

We need to optimize this case: calling scheduleLayerFlushNow() for invisible
pages. We do that by checking if the page is visible before scheduling
the RenderingUpdate.

Also we need to change or get rid of scheduleLayerFlushNow() since its name
has become confusing. It suggests that it is going to schedule flushing
the layer 'now'. But after r244182, it does scheduleRenderingUpdate() first.
And when it fires, scheduleCompositingLayerFlush() will be called.

* page/RenderingUpdateScheduler.cpp:
(WebCore::RenderingUpdateScheduler::scheduleRenderingUpdate):
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::scheduleLayerFlush):
(WebCore::RenderLayerCompositor::didChangeVisibleRect):
(WebCore::RenderLayerCompositor::frameViewDidScroll):
(WebCore::RenderLayerCompositor::attachRootLayer):
(WebCore::RenderLayerCompositor::setLayerFlushThrottlingEnabled):
(WebCore::RenderLayerCompositor::layerFlushTimerFired):
(WebCore::RenderLayerCompositor::scheduleLayerFlushNow): Deleted.
* rendering/RenderLayerCompositor.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244837 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoLink Previews that use WKImagePreviewViewController are not always scaled correctly
dino@apple.com [Wed, 1 May 2019 19:17:05 +0000 (19:17 +0000)]
Link Previews that use WKImagePreviewViewController are not always scaled correctly
https://bugs.webkit.org/show_bug.cgi?id=197450
<rdar://problem/50357695>

Reviewed by Wenson Hsieh.

Tell the UIImageView that we create in WKImagePreviewViewController to use
a filling scaling method.

* UIProcess/WKImagePreviewViewController.mm:
(-[WKImagePreviewViewController initWithCGImage:defaultActions:elementInfo:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244836 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoVersioning.
kocsen_chung@apple.com [Wed, 1 May 2019 18:26:13 +0000 (18:26 +0000)]
Versioning.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244835 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed correction to Test262 expectations following r244828.
ross.kirsling@sony.com [Wed, 1 May 2019 17:48:59 +0000 (17:48 +0000)]
Unreviewed correction to Test262 expectations following r244828.

* test262/expectations.yaml:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244833 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd memory-limited skipping to some tests generating very large strings
stephan.szabo@sony.com [Wed, 1 May 2019 17:42:51 +0000 (17:42 +0000)]
Add memory-limited skipping to some tests generating very large strings
https://bugs.webkit.org/show_bug.cgi?id=197437

Reviewed by Ross Kirsling.

* stress/StringObject-define-length-getter-rope-string-oom.js:
* stress/create-error-out-of-memory-rope-string.js:
* stress/string-16bit-repeat-overflow.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244832 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWebKit has too much of its own UTF-8 code and should rely more on ICU's UTF-8 support
ryanhaddad@apple.com [Wed, 1 May 2019 17:33:03 +0000 (17:33 +0000)]
WebKit has too much of its own UTF-8 code and should rely more on ICU's UTF-8 support
https://bugs.webkit.org/show_bug.cgi?id=195535

Patch by Darin Adler <darin@apple.com> on 2019-05-01
Reviewed by Alexey Proskuryakov.

LayoutTests/imported/w3c:

* web-platform-tests/encoding/textdecoder-utf16-surrogates-expected.txt:
Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8.

Source/JavaScriptCore:

* API/JSClassRef.cpp: Removed uneeded include of UTF8Conversion.h.

* API/JSStringRef.cpp:
(JSStringCreateWithUTF8CString): Updated for changes to convertUTF8ToUTF16.
(JSStringGetUTF8CString): Updated for changes to convertLatin1ToUTF8.
Removed unneeded "true" to get the strict version of convertUTF16ToUTF8,
since that is the default. Also updated for changes to CompletionResult.

* runtime/JSGlobalObjectFunctions.cpp:
(JSC::decode): Stop using UTF8SequenceLength, and instead use U8_COUNT_TRAIL_BYTES
and U8_MAX_LENGTH. Instead of decodeUTF8Sequence, use U8_NEXT. Also use U_IS_BMP,
U_IS_SUPPLEMENTARY, U16_LEAD, U16_TRAIL, and U_IS_SURROGATE instead of our own
equivalents, since these macros from ICU are correct and efficient.

* wasm/WasmParser.h:
(JSC::Wasm::Parser<SuccessType>::consumeUTF8String): Updated for changes to
convertUTF8ToUTF16.

Source/WebCore:

* platform/SharedBuffer.cpp:
(WebCore::utf8Buffer): Removed unnecessary "strict" argument to convertUTF16ToUTF8 since
that is the default behavior. Also updated for changes to return values.

* xml/XSLTProcessorLibxslt.cpp:
(WebCore::writeToStringBuilder): Removed unnecessary use of StringBuffer for a temporary
buffer for characters. Rewrote to use U8_NEXT and U16_APPEND directly.

* xml/parser/XMLDocumentParserLibxml2.cpp:
(WebCore::convertUTF16EntityToUTF8): Updated for changes to CompletionResult.

Source/WebKit:

* Shared/API/APIString.h: Removed uneeded includes and also switched to #pragma once.

* Shared/API/c/WKString.cpp: Moved include of UTF8Conversion.h here.
(WKStringGetUTF8CStringImpl): Updated for changes to return values.

Source/WTF:

* wtf/text/AtomicString.cpp:
(WTF::AtomicString::fromUTF8Internal): Added code to compute string length when the
end is nullptr; this behavior used to be implemented inside the
calculateStringHashAndLengthFromUTF8MaskingTop8Bits function.

* wtf/text/AtomicStringImpl.cpp:
(WTF::HashAndUTF8CharactersTranslator::translate): Updated for change to
convertUTF8ToUTF16.

* wtf/text/AtomicStringImpl.h: Took the WTF_EXPORT_PRIVATE off of the
AtomicStringImpl::addUTF8 function. This is used only inside a non-inlined function in
the AtomicString class and its behavior changed subtly in this patch; it's helpful
to document that it's not exported.

* wtf/text/StringImpl.cpp:
(WTF::StringImpl::utf8Impl): Don't pass "true" for strictness to convertUTF16ToUTF8
since strict is the default. Also updated for changes to ConversionResult.
(WTF::StringImpl::utf8ForCharacters): Updated for change to convertLatin1ToUTF8.
(WTF::StringImpl::tryGetUtf8ForRange const): Ditto.

* wtf/text/StringView.cpp: Removed uneeded include of UTF8Conversion.h.

* wtf/text/WTFString.cpp:
(WTF::String::fromUTF8): Updated for change to convertUTF8ToUTF16.

* wtf/unicode/UTF8Conversion.cpp:
(WTF::Unicode::inlineUTF8SequenceLengthNonASCII): Deleted.
(WTF::Unicode::inlineUTF8SequenceLength): Deleted.
(WTF::Unicode::UTF8SequenceLength): Deleted.
(WTF::Unicode::decodeUTF8Sequence): Deleted.
(WTF::Unicode::convertLatin1ToUTF8): Use U8_APPEND, enabling us to remove
almost everything in the function. Also changed resturn value to be a boolean
to indicate success since there is only one possible failure (target exhausted).
There is room for further simplification, since most callers have lengths rather
than end pointers for the source buffer, and all but one caller supplies a buffer
size known to be sufficient, so those don't need a return value, nor do they need
to pass an end of buffer pointer.
(WTF::Unicode::convertUTF16ToUTF8): Use U_IS_LEAD, U_IS_TRAIL,
U16_GET_SUPPLEMENTARY, U_IS_SURROGATE, and U8_APPEND. Also changed behavior
for non-strict mode so that unpaired surrogates will be turned into the
replacement character instead of invalid UTF-8 sequences, because U8_APPEND
won't create an invalid UTF-8 sequence, and because we don't need to do that
for any good reason at any call site.
(WTF::Unicode::isLegalUTF8): Deleted.
(WTF::Unicode::readUTF8Sequence): Deleted.
(WTF::Unicode::convertUTF8ToUTF16): Use U8_NEXT instead of
inlineUTF8SequenceLength, isLegalUTF8, and readUTF8Sequence. Use
U16_APPEND instead of lots of code that does the same thing. There is
room for further simplification since most callers don't need the "all ASCII"
feature and could probably pass the arguments in a more natural way.
(WTF::Unicode::calculateStringHashAndLengthFromUTF8MaskingTop8Bits):
Use U8_NEXT instead of isLegalUTF8, readUTF8Sequence, and various
error handling checks for things that are handled by U8_NEXT. Also removed
support for passing nullptr for end to specify a null-terminated string.
(WTF::Unicode::equalUTF16WithUTF8): Ditto.

* wtf/unicode/UTF8Conversion.h: Removed UTF8SequenceLength and
decodeUTF8Sequence. Changed the ConversionResult to match WebKit coding
style, with an eye toward perhaps removing it in the future. Changed
the convertUTF8ToUTF16 return value to a boolean and removed the "strict"
argument since no caller was passing false. Changed the convertLatin1ToUTF8
return value to a boolean. Tweaked comments.

LayoutTests:

* css3/escape-dom-api-expected.txt:
* fast/text/dangling-surrogates-expected.txt:
* js/dom/webidl-type-mapping-expected.txt:
* js/invalid-utf8-in-syntax-error-expected.txt:
Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8.

* js/invalid-utf8-in-syntax-error.html: Added. Before adding this, the test was
run, but unlike the rest of the tests in this directory, was only run as part of
run-javascriptcore-tests. There are two reasons for adding this. One is to be
consistent with the rest of the tests here and run a second time as part of the
broader WebKit tests. The second is that we can now use "--reset-results" to generate
new expected results, something that run-webkit-tests has but run-javascriptcore-tests
does not have.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244828 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244821.
sroberts@apple.com [Wed, 1 May 2019 17:13:58 +0000 (17:13 +0000)]
Unreviewed, rolling out r244821.

LayoutTests/imported/w3c:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/JavaScriptCore:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/WebCore:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/WebKit:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

Source/WTF:

Causing

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

LayoutTests:

Causing 4 Test262 failures on JSC Release and Debug

Reverted changeset:

"WebKit has too much of its own UTF-8 code and should rely
more on ICU's UTF-8 support"
https://bugs.webkit.org/show_bug.cgi?id=195535
https://trac.webkit.org/changeset/244821

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244827 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244822.
sroberts@apple.com [Wed, 1 May 2019 17:12:19 +0000 (17:12 +0000)]
Unreviewed, rolling out r244822.

Source/WebCore:

Causing

Reverted changeset:

https://trac.webkit.org/changeset/244822

Source/WTF:

Causing 4 Test262 failures on JSC Release and Debug

Reverted changeset:

https://trac.webkit.org/changeset/244822 https://trac.webkit.org/changeset/244821

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244826 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoReject/throw when calling AudioContext methods on a stopped AudioContext
youenn@apple.com [Wed, 1 May 2019 17:09:20 +0000 (17:09 +0000)]
Reject/throw when calling AudioContext methods on a stopped AudioContext
https://bugs.webkit.org/show_bug.cgi?id=197391

Reviewed by Eric Carlson.

Source/WebCore:

Return InvalidStateError in that case.
ASSERT that we do not call lazyInitialize after being stopped
since this would mean we are doing unneeded processing.

Test: http/wpt/webaudio/audiocontext-stopped.html

* Modules/webaudio/AudioContext.cpp:
(WebCore::AudioContext::lazyInitialize):
(WebCore::AudioContext::createBufferSource):
(WebCore::AudioContext::createMediaElementSource):
(WebCore::AudioContext::createMediaStreamSource):
(WebCore::AudioContext::createMediaStreamDestination):
(WebCore::AudioContext::createScriptProcessor):
(WebCore::AudioContext::createBiquadFilter):
(WebCore::AudioContext::createWaveShaper):
(WebCore::AudioContext::createPanner):
(WebCore::AudioContext::createConvolver):
(WebCore::AudioContext::createDynamicsCompressor):
(WebCore::AudioContext::createAnalyser):
(WebCore::AudioContext::createGain):
(WebCore::AudioContext::createDelay):
(WebCore::AudioContext::createChannelSplitter):
(WebCore::AudioContext::createChannelMerger):
(WebCore::AudioContext::createOscillator):
(WebCore::AudioContext::createPeriodicWave):
(WebCore::AudioContext::startRendering):
(WebCore::AudioContext::suspend):
(WebCore::AudioContext::resume):
(WebCore::AudioContext::close):
* Modules/webaudio/AudioContext.h:
* Modules/webaudio/AudioContext.idl:

LayoutTests:

* http/wpt/webaudio/audiocontext-stopped-expected.txt: Added.
* http/wpt/webaudio/audiocontext-stopped.html: Added.
* http/wpt/webaudio/resources/audiocontext-stopped-iframe.html: Added.
* platform/win/TestExpectations: Skip test for win.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244825 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoLayoutTests/imported/w3c:
youenn@apple.com [Wed, 1 May 2019 16:34:31 +0000 (16:34 +0000)]
LayoutTests/imported/w3c:
Kept alive loaders should use the redirected request in case of redirections
https://bugs.webkit.org/show_bug.cgi?id=197337

Reviewed by Alex Christensen.

* web-platform-tests/fetch/api/basic/keepalive-expected.txt:

Source/WebKit:
Kept alive loaders should use the redirected request in case of redirections
https://bugs.webkit.org/show_bug.cgi?id=197337

Reviewed by Alex Christensen.

Instead of using the request, the redirected request should be used in case of redirection.

* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::m_shouldCaptureExtraNetworkLoadMetrics):
(WebKit::NetworkResourceLoader::continueWillSendRedirectedRequest):

LayoutTests:
Enable Fetch Keep Alive by default
https://bugs.webkit.org/show_bug.cgi?id=197331

Reviewed by Alex Christensen.

* platform/mac-wk1/imported/w3c/web-platform-tests/fetch/api/basic/keepalive-expected.txt: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244824 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoXMLHttpRequest should propagate user gestures for media playback
eric.carlson@apple.com [Wed, 1 May 2019 15:56:34 +0000 (15:56 +0000)]
XMLHttpRequest should propagate user gestures for media playback
https://bugs.webkit.org/show_bug.cgi?id=197428
<rdar://problem/46677392>

Reviewed by Jer Noble.

Source/WebCore:

A user gesture the would allow media state change in effect when XMLHttpRequest.send is
called should be active when the event handlers fire after the transaction completes successfully.

Test: http/tests/media/user-gesture-preserved-across-xmlhttprequest.html

* dom/UserGestureIndicator.cpp:
(WebCore::UserGestureIndicator::UserGestureIndicator): Add a 'scope' parameter to potentially
limit the scope of the gesture to just media.
(WebCore::UserGestureIndicator::~UserGestureIndicator): Clear the scope.
* dom/UserGestureIndicator.h:
(WebCore::UserGestureToken::processingUserGesture const):
(WebCore::UserGestureToken::setScope):
(WebCore::UserGestureToken::resetScope):
(WebCore::UserGestureToken::hasExpired const):

* page/DOMTimer.cpp:
(WebCore::DOMTimerFireState::DOMTimerFireState): Don't need to store the nested timer interval,
UserGestureIndicator knows when it started.
(WebCore::DOMTimer::DOMTimer): Ditto.
(WebCore::DOMTimer::fired): Ditto.
(WebCore::DOMTimerFireState::nestedTimerInterval const): Deleted.
(WebCore::shouldForwardUserGesture): Deleted.
(WebCore::userGestureTokenToForward): Deleted.
(WebCore::currentNestedTimerInterval): Deleted.
* page/DOMTimer.h:

* testing/Internals.cpp:
(WebCore::Internals::setXHRMaximumIntervalForUserGestureForwarding): Override the maximum
user gesture interval for testing.
* testing/Internals.h:
* testing/Internals.idl:

* xml/XMLHttpRequest.cpp:
(WebCore::XMLHttpRequest::XMLHttpRequest):
(WebCore::XMLHttpRequest::send): Stash the user gesture token.
(WebCore::XMLHttpRequest::dispatchEvent): Clear user gesture token if it has expired. If still
valid, activate it.
* xml/XMLHttpRequest.h:

LayoutTests:

* fast/events/popup-blocking-timers5-expected.txt:
* fast/events/popup-blocking-timers5.html: Decrease the timer interval from 1000 to 900
because the user gesture is invalidated based on wall clock time.
* fast/events/popup-blocking-timers6-expected.txt:
* fast/events/popup-blocking-timers6.html: Increase the timer interval from 1001 to 1100
because the user gesture is invalidated based on wall clock time.
* http/tests/media/user-gesture-preserved-across-xmlhttprequest-expected.txt: Added.
* http/tests/media/user-gesture-preserved-across-xmlhttprequest.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244823 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFix typos: "convering" to "converting".
darin@apple.com [Wed, 1 May 2019 15:55:10 +0000 (15:55 +0000)]
Fix typos: "convering" to "converting".

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244822 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWebKit has too much of its own UTF-8 code and should rely more on ICU's UTF-8 support
darin@apple.com [Wed, 1 May 2019 15:52:16 +0000 (15:52 +0000)]
WebKit has too much of its own UTF-8 code and should rely more on ICU's UTF-8 support
https://bugs.webkit.org/show_bug.cgi?id=195535

Reviewed by Alexey Proskuryakov.

LayoutTests/imported/w3c:

* web-platform-tests/encoding/textdecoder-utf16-surrogates-expected.txt:
Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8.

Source/JavaScriptCore:

* API/JSClassRef.cpp: Removed uneeded include of UTF8Conversion.h.

* API/JSStringRef.cpp:
(JSStringCreateWithUTF8CString): Updated for changes to convertUTF8ToUTF16.
(JSStringGetUTF8CString): Updated for changes to convertLatin1ToUTF8.
Removed unneeded "true" to get the strict version of convertUTF16ToUTF8,
since that is the default. Also updated for changes to CompletionResult.

* runtime/JSGlobalObjectFunctions.cpp:
(JSC::decode): Stop using UTF8SequenceLength, and instead use U8_COUNT_TRAIL_BYTES
and U8_MAX_LENGTH. Instead of decodeUTF8Sequence, use U8_NEXT. Also use U_IS_BMP,
U_IS_SUPPLEMENTARY, U16_LEAD, U16_TRAIL, and U_IS_SURROGATE instead of our own
equivalents, since these macros from ICU are correct and efficient.

* wasm/WasmParser.h:
(JSC::Wasm::Parser<SuccessType>::consumeUTF8String): Updated for changes to
convertUTF8ToUTF16.

Source/WebCore:

* platform/SharedBuffer.cpp:
(WebCore::utf8Buffer): Removed unnecessary "strict" argument to convertUTF16ToUTF8 since
that is the default behavior. Also updated for changes to return values.

* xml/XSLTProcessorLibxslt.cpp:
(WebCore::writeToStringBuilder): Removed unnecessary use of StringBuffer for a temporary
buffer for characters. Rewrote to use U8_NEXT and U16_APPEND directly.

* xml/parser/XMLDocumentParserLibxml2.cpp:
(WebCore::convertUTF16EntityToUTF8): Updated for changes to CompletionResult.

Source/WebKit:

* Shared/API/APIString.h: Removed uneeded includes and also switched to #pragma once.

* Shared/API/c/WKString.cpp: Moved include of UTF8Conversion.h here.
(WKStringGetUTF8CStringImpl): Updated for changes to return values.

Source/WTF:

* wtf/text/AtomicString.cpp:
(WTF::AtomicString::fromUTF8Internal): Added code to compute string length when the
end is nullptr; this behavior used to be implemented inside the
calculateStringHashAndLengthFromUTF8MaskingTop8Bits function.

* wtf/text/AtomicStringImpl.cpp:
(WTF::HashAndUTF8CharactersTranslator::translate): Updated for change to
convertUTF8ToUTF16.

* wtf/text/AtomicStringImpl.h: Took the WTF_EXPORT_PRIVATE off of the
AtomicStringImpl::addUTF8 function. This is used only inside a non-inlined function in
the AtomicString class and its behavior changed subtly in this patch; it's helpful
to document that it's not exported.

* wtf/text/StringImpl.cpp:
(WTF::StringImpl::utf8Impl): Don't pass "true" for strictness to convertUTF16ToUTF8
since strict is the default. Also updated for changes to ConversionResult.
(WTF::StringImpl::utf8ForCharacters): Updated for change to convertLatin1ToUTF8.
(WTF::StringImpl::tryGetUtf8ForRange const): Ditto.

* wtf/text/StringView.cpp: Removed uneeded include of UTF8Conversion.h.

* wtf/text/WTFString.cpp:
(WTF::String::fromUTF8): Updated for change to convertUTF8ToUTF16.

* wtf/unicode/UTF8Conversion.cpp:
(WTF::Unicode::inlineUTF8SequenceLengthNonASCII): Deleted.
(WTF::Unicode::inlineUTF8SequenceLength): Deleted.
(WTF::Unicode::UTF8SequenceLength): Deleted.
(WTF::Unicode::decodeUTF8Sequence): Deleted.
(WTF::Unicode::convertLatin1ToUTF8): Use U8_APPEND, enabling us to remove
almost everything in the function. Also changed resturn value to be a boolean
to indicate success since there is only one possible failure (target exhausted).
There is room for further simplification, since most callers have lengths rather
than end pointers for the source buffer, and all but one caller supplies a buffer
size known to be sufficient, so those don't need a return value, nor do they need
to pass an end of buffer pointer.
(WTF::Unicode::convertUTF16ToUTF8): Use U_IS_LEAD, U_IS_TRAIL,
U16_GET_SUPPLEMENTARY, U_IS_SURROGATE, and U8_APPEND. Also changed behavior
for non-strict mode so that unpaired surrogates will be turned into the
replacement character instead of invalid UTF-8 sequences, because U8_APPEND
won't create an invalid UTF-8 sequence, and because we don't need to do that
for any good reason at any call site.
(WTF::Unicode::isLegalUTF8): Deleted.
(WTF::Unicode::readUTF8Sequence): Deleted.
(WTF::Unicode::convertUTF8ToUTF16): Use U8_NEXT instead of
inlineUTF8SequenceLength, isLegalUTF8, and readUTF8Sequence. Use
U16_APPEND instead of lots of code that does the same thing. There is
room for further simplification since most callers don't need the "all ASCII"
feature and could probably pass the arguments in a more natural way.
(WTF::Unicode::calculateStringHashAndLengthFromUTF8MaskingTop8Bits):
Use U8_NEXT instead of isLegalUTF8, readUTF8Sequence, and various
error handling checks for things that are handled by U8_NEXT. Also removed
support for passing nullptr for end to specify a null-terminated string.
(WTF::Unicode::equalUTF16WithUTF8): Ditto.

* wtf/unicode/UTF8Conversion.h: Removed UTF8SequenceLength and
decodeUTF8Sequence. Changed the ConversionResult to match WebKit coding
style, with an eye toward perhaps removing it in the future. Changed
the convertUTF8ToUTF16 return value to a boolean and removed the "strict"
argument since no caller was passing false. Changed the convertLatin1ToUTF8
return value to a boolean. Tweaked comments.

LayoutTests:

* css3/escape-dom-api-expected.txt:
* fast/text/dangling-surrogates-expected.txt:
* js/dom/webidl-type-mapping-expected.txt:
* js/invalid-utf8-in-syntax-error-expected.txt:
Updated expected results to have the Unicode replacement character in cases where the
text contains unpaired surrogates. The tests are still doing the same operations, and
still getting the same results, but the text output no longer includes illegal UTF-8
because the WTF changes affect the code path that DumpRenderTree and WebKitTestRunner
use to produce the text output.

* js/invalid-utf8-in-syntax-error.html: Added. Before adding this, the test was
run, but unlike the rest of the tests in this directory, was only run as part of
run-javascriptcore-tests. There are two reasons for adding this. One is to be
consistent with the rest of the tests here and run a second time as part of the
broader WebKit tests. The second is that we can now use "--reset-results" to generate
new expected results, something that run-webkit-tests has but run-javascriptcore-tests
does not have.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244821 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFix the WebKitTestRunner build
timothy_horton@apple.com [Wed, 1 May 2019 06:30:51 +0000 (06:30 +0000)]
Fix the WebKitTestRunner build
https://bugs.webkit.org/show_bug.cgi?id=197449
<rdar://problem/50334169>

Reviewed by Alexey Proskuryakov.

* WebKitTestRunner/mac/EventSenderProxy.mm:
(-[EventSenderSyntheticEvent initPressureEventAtLocation:globalLocation:stage:pressure:stageTransition:phase:time:eventNumber:window:]):
(WTR::EventSenderProxy::mouseForceChanged):
Use some SPI instead of IPI.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244820 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRegression(PSON) URL scheme handlers can no longer respond asynchronously
cdumez@apple.com [Wed, 1 May 2019 02:45:10 +0000 (02:45 +0000)]
Regression(PSON) URL scheme handlers can no longer respond asynchronously
https://bugs.webkit.org/show_bug.cgi?id=197426
<rdar://problem/50256169>

Reviewed by Brady Eidson.

Source/WebKit:

The issue was that when committing the provisional process, we would call WebPageProxy::processDidTerminate()
which would call WebPageProxy::stopAllURLSchemeTasks(). This would terminate all URL scheme tasks associated
with the page, including the one associated with the provisisional page / process.

To address the issue, pass m_process to stopAllURLSchemeTasks() in processDidTerminate() and only stop the
tasks associated with the m_process (which is the process we're about to swap away from).

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::processDidTerminate):
(WebKit::WebPageProxy::stopAllURLSchemeTasks):
* UIProcess/WebPageProxy.h:
* UIProcess/WebURLSchemeHandler.cpp:
(WebKit::WebURLSchemeHandler::processForTaskIdentifier):
(WebKit::WebURLSchemeHandler::stopAllTasksForPage):
* UIProcess/WebURLSchemeHandler.h:
* UIProcess/WebURLSchemeTask.h:
(WebKit::WebURLSchemeTask::process const):

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
(-[PSONScheme setShouldRespondAsynchronously:]):
(-[PSONScheme webView:startURLSchemeTask:]):
(-[PSONScheme webView:stopURLSchemeTask:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244819 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd logging of Ad Click Attribution errors and events to a dedicated channel
wilander@apple.com [Wed, 1 May 2019 02:28:07 +0000 (02:28 +0000)]
Add logging of Ad Click Attribution errors and events to a dedicated channel
https://bugs.webkit.org/show_bug.cgi?id=197332
<rdar://problem/49918800>

Reviewed by Youenn Fablet.

Source/WebCore:

This patch adds an experimental Ad Click Attribution debug mode which
logs information.

No new tests.

* loader/AdClickAttribution.cpp:
(WebCore::AdClickAttribution::parseConversionRequest):
(WebCore::AdClickAttribution::debugModeEnabled):
* loader/AdClickAttribution.h:
* page/RuntimeEnabledFeatures.h:
(WebCore::RuntimeEnabledFeatures::adClickAttributionDebugModeEnabled const):
(WebCore::RuntimeEnabledFeatures::setAdClickAttributionDebugModeEnabled):
* platform/Logging.h:

Source/WebKit:

This patch adds an experimental Ad Click Attribution debug mode which
logs information. Most changes are just log output in the various
functions in WebKit::AdClickAttributionManager.

The constructor to WebKit::AdClickAttributionManager now takes a
PAL::SessionID so that the log functions can make sure they don't
output anything in ephemeral sessions.

WebProcessPool::platformInitializeNetworkProcess() now picks up the
debug mode setting from the incoming
WebKit::NetworkProcessCreationParameters object.

NetworkResourceLoader::handleAdClickAttributionConversion() was
moved to AdClickAttributionManager::handleConversion() to keep all
the logging in one file.

* NetworkProcess/AdClickAttributionManager.cpp:
(WebKit::AdClickAttributionManager::storeUnconverted):
(WebKit::AdClickAttributionManager::handleConversion):
(WebKit::AdClickAttributionManager::convert):
(WebKit::AdClickAttributionManager::fireConversionRequest):
(WebKit::AdClickAttributionManager::firePendingConversionRequests):
(WebKit::AdClickAttributionManager::clearExpired):
(WebKit::AdClickAttributionManager::debugModeEnabled const):
* NetworkProcess/AdClickAttributionManager.h:
(WebKit::AdClickAttributionManager::AdClickAttributionManager):
(WebKit::AdClickAttributionManager::m_sessionID):
* NetworkProcess/NetworkProcess.cpp:
(WebKit::NetworkProcess::initializeNetworkProcess):
* NetworkProcess/NetworkProcessCreationParameters.cpp:
(WebKit::NetworkProcessCreationParameters::encode const):
(WebKit::NetworkProcessCreationParameters::decode):
* NetworkProcess/NetworkProcessCreationParameters.h:
* NetworkProcess/NetworkResourceLoader.cpp:
(WebKit::NetworkResourceLoader::continueWillSendRedirectedRequest):
(WebKit::NetworkResourceLoader::handleAdClickAttributionConversion): Deleted.
* NetworkProcess/NetworkSession.cpp:
(WebKit::NetworkSession::NetworkSession):
(WebKit::NetworkSession::handleAdClickAttributionConversion):
(WebKit::NetworkSession::convertAdClickAttribution): Deleted.
* NetworkProcess/NetworkSession.h:
* Platform/Logging.h:
* Shared/WebPreferences.yaml:
* UIProcess/Cocoa/WebProcessPoolCocoa.mm:
(WebKit::WebProcessPool::platformInitializeNetworkProcess):

Source/WTF:

Added missing RELEASE_LOG_INFO and RELEASE_LOG_INFO_IF dummies
for RELEASE_LOG_DISABLED.

* wtf/Assertions.h:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244818 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agofont-weight: 1000 is not parsed successfully
mmaxfield@apple.com [Wed, 1 May 2019 01:14:37 +0000 (01:14 +0000)]
font-weight: 1000 is not parsed successfully
https://bugs.webkit.org/show_bug.cgi?id=197427

Reviewed by Dean Jackson.

Source/WebCore:

The spec says:
"Only values greater than or equal to 1, and less than or equal to 1000, are valid"

This change brings us in-line with all the other browsers.

Test: fast/text/font-weight-1-1000.html

* css/parser/CSSPropertyParserHelpers.cpp:
(WebCore::CSSPropertyParserHelpers::consumeFontWeightNumber):

LayoutTests:

* platform/win/TestExpectations: Windows disables variation fonts, so requires values to be divisible by 100
* fast/text/font-weight-1-1000-expected.txt: Added.
* fast/text/font-weight-1-1000.html: Added.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244817 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244806.
commit-queue@webkit.org [Wed, 1 May 2019 00:32:15 +0000 (00:32 +0000)]
Unreviewed, rolling out r244806.
https://bugs.webkit.org/show_bug.cgi?id=197446

Causing Test262 and JSC test failures on multiple builds
(Requested by ShawnRoberts on #webkit).

Reverted changeset:

"TypeArrays should not store properties that are canonical
numeric indices"
https://bugs.webkit.org/show_bug.cgi?id=197228
https://trac.webkit.org/changeset/244806

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244816 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoMake Document audio producers use WeakPtr
youenn@apple.com [Wed, 1 May 2019 00:13:08 +0000 (00:13 +0000)]
Make Document audio producers use WeakPtr
https://bugs.webkit.org/show_bug.cgi?id=197382

Reviewed by Eric Carlson.

Source/WebCore:

Move from a hash set of raw pointers to a hash set of weak pointers.
This helps make the code cleaner.
No observable change of behavior.

* Modules/mediastream/MediaStreamTrack.h:
* dom/Document.cpp:
(WebCore::Document::addAudioProducer):
(WebCore::Document::removeAudioProducer):
(WebCore::Document::updateIsPlayingMedia):
(WebCore::Document::pageMutedStateDidChange):
* dom/Document.h:
* html/HTMLMediaElement.cpp:
(WebCore::HTMLMediaElement::updateActiveTextTrackCues):
* html/HTMLMediaElement.h:
* page/MediaProducer.h:

Source/WTF:

* wtf/WeakHashSet.h:
(WTF::WeakHashSet::hasNullReferences const):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244815 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoCrash when running test wpt/tests/element_click/bubbling.py::test_element_disappears_...
drousso@apple.com [Wed, 1 May 2019 00:12:51 +0000 (00:12 +0000)]
Crash when running test wpt/tests/element_click/bubbling.py::test_element_disappears_during_click
https://bugs.webkit.org/show_bug.cgi?id=197361
<rdar://problem/49861407>

Reviewed by Brian Burg.

Don't assume that all elements have client rects (e.g. `getClientRects`). If the container
element isn't visible, then the child won't be either.

* UIProcess/Automation/SimulatedInputDispatcher.cpp:
(WebKit::SimulatedInputDispatcher::resolveLocation):
(WebKit::SimulatedInputDispatcher::transitionInputSourceToState):
* UIProcess/Automation/WebAutomationSession.cpp:
(WebKit::WebAutomationSession::viewportInViewCenterPointOfElement):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244814 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[macOS WK1] ASSERTION FAILED: formData in WebCore::ResourceRequest::doUpdateResourceH...
youenn@apple.com [Tue, 30 Apr 2019 23:56:20 +0000 (23:56 +0000)]
[macOS WK1] ASSERTION FAILED: formData in WebCore::ResourceRequest::doUpdateResourceHTTPBody()
https://bugs.webkit.org/show_bug.cgi?id=196864
<rdar://problem/49854497>

Reviewed by Alex Christensen.

LayoutTests/imported/w3c:

* web-platform-tests/xhr/send-redirect-post-upload-expected.txt:

Source/WebCore:

In case of redirection, it is sometimes not possible to retrieve the form data
from its NSInputStream in case of redirections.
To handle this case, reuse the first request form data if the new request has a body.
We also clear the HTTP content type in such a case if the original request has no content type.

Covered by re-enabled tests.

* platform/network/mac/WebCoreResourceHandleAsOperationQueueDelegate.mm:
(-[WebCoreResourceHandleAsOperationQueueDelegate connection:willSendRequest:redirectResponse:]):

LayoutTests:

* platform/mac-wk1/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244813 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244802.
cdumez@apple.com [Tue, 30 Apr 2019 23:56:05 +0000 (23:56 +0000)]
Unreviewed, rolling out r244802.

Caused an API test failure

Reverted changeset:

"Regression(PSON) URL scheme handlers can no longer respond
asynchronously"
https://bugs.webkit.org/show_bug.cgi?id=197426
https://trac.webkit.org/changeset/244802

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244812 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoCodeBlock::m_instructionCount is wrong
sbarati@apple.com [Tue, 30 Apr 2019 23:37:27 +0000 (23:37 +0000)]
CodeBlock::m_instructionCount is wrong
https://bugs.webkit.org/show_bug.cgi?id=197304

Reviewed by Yusuke Suzuki.

What we were calling instructionCount() was wrong, as evidenced by
us using it incorrectly both in the sampling profiler and when we
dumped bytecode for a given CodeBlock. Prior to the bytecode rewrite,
instructionCount() was probably valid to do bounds checks against.
However, this is no longer the case. This patch renames what we called
instructionCount() to bytecodeCost(). It is now only used to make decisions
about inlining and tier up heuristics. I've also named options related to
this appropriately.

This patch also introduces instructionsSize(). The result of this method
is valid to do bounds checks against.

* bytecode/CodeBlock.cpp:
(JSC::CodeBlock::dumpAssumingJITType const):
(JSC::CodeBlock::CodeBlock):
(JSC::CodeBlock::finishCreation):
(JSC::CodeBlock::optimizationThresholdScalingFactor):
(JSC::CodeBlock::predictedMachineCodeSize):
* bytecode/CodeBlock.h:
(JSC::CodeBlock::instructionsSize const):
(JSC::CodeBlock::bytecodeCost const):
(JSC::CodeBlock::instructionCount const): Deleted.
* dfg/DFGByteCodeParser.cpp:
(JSC::DFG::ByteCodeParser::inliningCost):
(JSC::DFG::ByteCodeParser::getInliningBalance):
* dfg/DFGCapabilities.cpp:
(JSC::DFG::mightCompileEval):
(JSC::DFG::mightCompileProgram):
(JSC::DFG::mightCompileFunctionForCall):
(JSC::DFG::mightCompileFunctionForConstruct):
(JSC::DFG::mightInlineFunctionForCall):
(JSC::DFG::mightInlineFunctionForClosureCall):
(JSC::DFG::mightInlineFunctionForConstruct):
* dfg/DFGCapabilities.h:
(JSC::DFG::isSmallEnoughToInlineCodeInto):
* dfg/DFGDisassembler.cpp:
(JSC::DFG::Disassembler::dumpHeader):
* dfg/DFGDriver.cpp:
(JSC::DFG::compileImpl):
* dfg/DFGPlan.cpp:
(JSC::DFG::Plan::compileInThread):
* dfg/DFGTierUpCheckInjectionPhase.cpp:
(JSC::DFG::TierUpCheckInjectionPhase::run):
* ftl/FTLCapabilities.cpp:
(JSC::FTL::canCompile):
* ftl/FTLCompile.cpp:
(JSC::FTL::compile):
* ftl/FTLLink.cpp:
(JSC::FTL::link):
* jit/JIT.cpp:
(JSC::JIT::link):
* jit/JITDisassembler.cpp:
(JSC::JITDisassembler::dumpHeader):
* llint/LLIntSlowPaths.cpp:
(JSC::LLInt::shouldJIT):
* profiler/ProfilerBytecodes.cpp:
(JSC::Profiler::Bytecodes::Bytecodes):
* runtime/Options.h:
* runtime/SamplingProfiler.cpp:
(JSC::tryGetBytecodeIndex):
(JSC::SamplingProfiler::processUnverifiedStackTraces):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244811 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoimported/w3c/web-platform-tests/xhr/event-upload-progress-crossorigin.htm is a flaky...
sroberts@apple.com [Tue, 30 Apr 2019 22:49:04 +0000 (22:49 +0000)]
imported/w3c/web-platform-tests/xhr/event-upload-progress-crossorigin.htm is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=182849

Unreviewed test gardening.

* platform/ios-simulator-wk2/TestExpectations: Updating expectations for flaky failure on iOS

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244808 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoTypeArrays should not store properties that are canonical numeric indices
tzagallo@apple.com [Tue, 30 Apr 2019 22:25:09 +0000 (22:25 +0000)]
TypeArrays should not store properties that are canonical numeric indices
https://bugs.webkit.org/show_bug.cgi?id=197228
<rdar://problem/49557381>

Reviewed by Darin Adler.

JSTests:

* stress/typed-array-canonical-numeric-index-string.js: Added.
(makeTest.assert):
(makeTest):
(const.testInvalidIndices.makeTest.set assert):
(const.testInvalidIndices.makeTest):
(const.testValidIndices.makeTest.set assert):
(const.testValidIndices.makeTest):

Source/JavaScriptCore:

According to the spec[1], TypedArrays should not perform an ordinary GetOwnProperty/SetOwnProperty
if the index is a CanonicalNumericIndexString, but invalid according toIntegerIndexedElementGet
and similar functions. I.e., there are a few properties that should not be set in a TypedArray,
like NaN, Infinity and -0.

[1]: https://www.ecma-international.org/ecma-262/9.0/index.html#sec-integer-indexed-exotic-objects-defineownproperty-p-desc

* CMakeLists.txt:
* JavaScriptCore.xcodeproj/project.pbxproj:
* runtime/JSGenericTypedArrayViewInlines.h:
(JSC::JSGenericTypedArrayView<Adaptor>::getOwnPropertySlot):
(JSC::JSGenericTypedArrayView<Adaptor>::put):
(JSC::JSGenericTypedArrayView<Adaptor>::defineOwnProperty):
(JSC::JSGenericTypedArrayView<Adaptor>::getOwnPropertySlotByIndex):
(JSC::JSGenericTypedArrayView<Adaptor>::putByIndex):
* runtime/JSTypedArrays.cpp:
* runtime/PropertyName.h:
(JSC::canonicalNumericIndexString):

LayoutTests:

* fast/canvas/canvas-ImageData-behaviour-expected.txt:
* fast/canvas/canvas-ImageData-behaviour.js:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244806 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUpdate Spanish Translation
mcatanzaro@igalia.com [Tue, 30 Apr 2019 22:12:11 +0000 (22:12 +0000)]
Update Spanish Translation
https://bugs.webkit.org/show_bug.cgi?id=197407

Patch by Álvaro Torralba <donfrutosgomez@gmail.com> on 2019-04-30
Rubber-stamped by Michael Catanzaro.

* es.po:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244805 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoRegression(PSON) URL scheme handlers can no longer respond asynchronously
cdumez@apple.com [Tue, 30 Apr 2019 21:11:07 +0000 (21:11 +0000)]
Regression(PSON) URL scheme handlers can no longer respond asynchronously
https://bugs.webkit.org/show_bug.cgi?id=197426
<rdar://problem/50256169>

Reviewed by Brady Eidson.

Source/WebKit:

The issue was that when committing the provisional process, we would call WebPageProxy::processDidTerminate()
which would call WebPageProxy::stopAllURLSchemeTasks(). This would terminate all URL scheme tasks associated
with the page, including the one associated with the provisisional page / process.

To address the issue, pass m_process to stopAllURLSchemeTasks() in processDidTerminate() and only stop the
tasks associated with the m_process (which is the process we're about to swap away from).

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::processDidTerminate):
(WebKit::WebPageProxy::stopAllURLSchemeTasks):
* UIProcess/WebPageProxy.h:
* UIProcess/WebURLSchemeHandler.cpp:
(WebKit::WebURLSchemeHandler::processForTaskIdentifier):
(WebKit::WebURLSchemeHandler::stopAllTasksForPage):
* UIProcess/WebURLSchemeHandler.h:
* UIProcess/WebURLSchemeTask.h:
(WebKit::WebURLSchemeTask::process const):

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:
(-[PSONScheme setShouldRespondAsynchronously:]):
(-[PSONScheme webView:startURLSchemeTask:]):
(-[PSONScheme webView:stopURLSchemeTask:]):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244802 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244773.
commit-queue@webkit.org [Tue, 30 Apr 2019 20:21:11 +0000 (20:21 +0000)]
Unreviewed, rolling out r244773.
https://bugs.webkit.org/show_bug.cgi?id=197436

Causing assertion failures on debug queues (Requested by
ShawnRoberts on #webkit).

Reverted changeset:

"Make Document audio producers use WeakPtr"
https://bugs.webkit.org/show_bug.cgi?id=197382
https://trac.webkit.org/changeset/244773

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244801 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoTransform is sometimes left in a bad state after an animation
simon.fraser@apple.com [Tue, 30 Apr 2019 20:20:37 +0000 (20:20 +0000)]
Transform is sometimes left in a bad state after an animation
https://bugs.webkit.org/show_bug.cgi?id=197401
rdar://problem/48179186

Reviewed by Dean Jackson.

Source/WebCore:

In some more complex compositing scenarios, at the end of an animation we'd
fail to push a new transform onto a layer, because updateGeometry() would
think there's an animation running (which there is, but in the "Ending" state).

It's simpler in this code to just always push transform and opacity to the layer;
they will get overridden by the animation while it's running. The current code
dates from the first landing of the file, and the reason for the if (!isRunningAcceleratedTransformAnimation)
check is lost in the sands of time.

I was not able to get a reliable ref or layer tree test for this, because the next compositing update
fixes it, and WTR seems to trigger one.  But the added test does show the bug
in Safari, and is a good test to have.

Test: compositing/animation/transform-after-animation.html

* rendering/RenderLayerBacking.cpp:
(WebCore::RenderLayerBacking::updateGeometry):

LayoutTests:

Share code between all the overlap tests that work by creating dot matrices, and strip
out the transforms from the layer tree dumps, because they can vary with timing in these
tests.

* compositing/animation/transform-after-animation-expected.html: Added.
* compositing/animation/transform-after-animation.html: Added.
* compositing/backing/backing-store-attachment-empty-keyframe-expected.txt:
* compositing/layer-creation/animation-overlap-with-children.html:
* compositing/layer-creation/mismatched-rotated-transform-animation-overlap-expected.txt:
* compositing/layer-creation/mismatched-rotated-transform-animation-overlap.html:
* compositing/layer-creation/mismatched-rotated-transform-transition-overlap.html:
* compositing/layer-creation/mismatched-transform-transition-overlap.html:
* compositing/layer-creation/multiple-keyframes-animation-overlap-expected.txt:
* compositing/layer-creation/multiple-keyframes-animation-overlap.html:
* compositing/layer-creation/resources/compositing-overlap-utils.js: Added.
(makeDots):
(layerTreeWithoutTransforms):
(dumpLayers):
* compositing/layer-creation/scale-rotation-animation-overlap-expected.txt:
* compositing/layer-creation/scale-rotation-animation-overlap.html:
* compositing/layer-creation/scale-rotation-transition-overlap.html:
* compositing/layer-creation/translate-animation-overlap-expected.txt:
* compositing/layer-creation/translate-animation-overlap.html:
* compositing/layer-creation/translate-scale-animation-overlap-expected.txt:
* compositing/layer-creation/translate-scale-animation-overlap.html:
* compositing/layer-creation/translate-scale-transition-overlap.html:
* compositing/layer-creation/translate-transition-overlap.html:
* legacy-animation-engine/compositing/backing/backing-store-attachment-empty-keyframe-expected.txt:
* legacy-animation-engine/compositing/backing/backing-store-attachment-empty-keyframe.html:
* legacy-animation-engine/compositing/backing/transform-transition-from-outside-view-expected.txt:
* legacy-animation-engine/compositing/layer-creation/animation-overlap-with-children.html:
* legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-animation-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-animation-overlap.html:
* legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-transition-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/mismatched-rotated-transform-transition-overlap.html:
* legacy-animation-engine/compositing/layer-creation/mismatched-transform-transition-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/mismatched-transform-transition-overlap.html:
* legacy-animation-engine/compositing/layer-creation/multiple-keyframes-animation-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/multiple-keyframes-animation-overlap.html:
* legacy-animation-engine/compositing/layer-creation/scale-rotation-animation-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/scale-rotation-animation-overlap.html:
* legacy-animation-engine/compositing/layer-creation/scale-rotation-transition-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/scale-rotation-transition-overlap.html:
* legacy-animation-engine/compositing/layer-creation/translate-animation-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/translate-animation-overlap.html:
* legacy-animation-engine/compositing/layer-creation/translate-scale-animation-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/translate-scale-animation-overlap.html:
* legacy-animation-engine/compositing/layer-creation/translate-scale-transition-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/translate-scale-transition-overlap.html:
* legacy-animation-engine/compositing/layer-creation/translate-transition-overlap-expected.txt:
* legacy-animation-engine/compositing/layer-creation/translate-transition-overlap.html:
* platform/ios/compositing/layer-creation/scale-rotation-animation-overlap-expected.txt:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244800 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoOnly use a related page's process if that page has not been closed yet
cdumez@apple.com [Tue, 30 Apr 2019 20:20:00 +0000 (20:20 +0000)]
Only use a related page's process if that page has not been closed yet
https://bugs.webkit.org/show_bug.cgi?id=197393
<rdar://problem/50302423>

Reviewed by Tim Horton.

Source/WebKit:

We should not attempt to use a related page's process if that related page has already been closed.
Once closed, a page's process is invalid and trying to launch a new process for the closed page
leads to crashes such as the one in the radar.

* UIProcess/WebPageProxy.cpp:
(WebKit::WebPageProxy::launchProcess):
* UIProcess/WebProcessPool.cpp:
(WebKit::WebProcessPool::createWebPage):

Tools:

Add API test coverage.

* TestWebKitAPI/Tests/WebKitCocoa/ProcessSwapOnNavigation.mm:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244799 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoUnreviewed, rolling out r244774.
commit-queue@webkit.org [Tue, 30 Apr 2019 20:09:44 +0000 (20:09 +0000)]
Unreviewed, rolling out r244774.
https://bugs.webkit.org/show_bug.cgi?id=197431

Causing assertion failures on debug queues (Requested by
ShawnRoberts on #webkit).

Reverted changeset:

"Reject/throw when calling AudioContext methods on a stopped
AudioContext"
https://bugs.webkit.org/show_bug.cgi?id=197391
https://trac.webkit.org/changeset/244774

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244797 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoAdd WKContentRuleList ping resource-type
achristensen@apple.com [Tue, 30 Apr 2019 19:42:34 +0000 (19:42 +0000)]
Add WKContentRuleList ping resource-type
https://bugs.webkit.org/show_bug.cgi?id=197325
<rdar://problem/49841404>

Reviewed by Geoff Garen.

Source/WebCore:

Tests: http/tests/contentextensions/block-ping-resource-type-ping.html and http/tests/contentextensions/block-ping-resource-type-raw.html

* contentextensions/ContentExtensionsBackend.cpp:
(WebCore::ContentExtensions::ContentExtensionsBackend::processContentRuleListsForLoad):
* contentextensions/ContentExtensionsBackend.h:
* loader/PingLoader.cpp:
(WebCore::processContentRuleListsForLoad):
(WebCore::PingLoader::sendPing):
* loader/ResourceLoadInfo.cpp:
(WebCore::ContentExtensions::readResourceType):
(WebCore::ContentExtensions::ResourceLoadInfo::getResourceFlags const):
* loader/ResourceLoadInfo.h:
* page/UserContentProvider.cpp:
(WebCore::UserContentProvider::processContentRuleListsForLoad):
* page/UserContentProvider.h:

LayoutTests:

* http/tests/contentextensions/block-ping-resource-type-raw--expected.txt: Copied from LayoutTests/http/tests/contentextensions/block-ping-expected.txt.
* http/tests/contentextensions/block-ping-resource-type-raw.html: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.
* http/tests/contentextensions/block-ping-resource-type-raw.html.json: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.json.
* http/tests/contentextensions/block-ping-resource-type-ping-expected.txt: Copied from LayoutTests/http/tests/contentextensions/block-ping-expected.txt.
* http/tests/contentextensions/block-ping-resource-type-ping.html: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.
* http/tests/contentextensions/block-ping-resource-type-ping.html.json: Copied from LayoutTests/http/tests/contentextensions/block-ping.html.json.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244794 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Enable Bindings tests queue on new EWS
aakash_jain@apple.com [Tue, 30 Apr 2019 18:42:38 +0000 (18:42 +0000)]
[ews-build] Enable Bindings tests queue on new EWS
https://bugs.webkit.org/show_bug.cgi?id=197424

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-app/ews/views/statusbubble.py:
(StatusBubble):
* BuildSlaveSupport/ews-build/config.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244792 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoWeb Automation: use a more informative key to indicate automation availability
bburg@apple.com [Tue, 30 Apr 2019 18:38:16 +0000 (18:38 +0000)]
Web Automation: use a more informative key to indicate automation availability
https://bugs.webkit.org/show_bug.cgi?id=197377
<rdar://problem/50258069>

Reviewed by Devin Rousso.

The existing WIRAutomationEnabledKey does not encode uncertainty.
Add a new key that provides an 'Unknown' state, and prefer to use it.

Since an application's initial listing is sent from a background dispatch queue
on Cocoa platforms, this can race with main thread initialization that sets up
RemoteInspector::Client. Therefore, the initial listing may not properly represent
the client's capabilites because the client is not yet available. Allowing for
an "Unknown" state that is later upgraded to Available or Not Available makes it
possible to work around this potential race.

* inspector/remote/RemoteInspectorConstants.h:
* inspector/remote/cocoa/RemoteInspectorCocoa.mm:
(Inspector::RemoteInspector::pushListingsNow):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244791 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Configure ews151 for running various tests
aakash_jain@apple.com [Tue, 30 Apr 2019 18:35:03 +0000 (18:35 +0000)]
[ews-build] Configure ews151 for running various tests
https://bugs.webkit.org/show_bug.cgi?id=197419

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/config.json:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244790 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[ews-build] Parse and display bindings test failures
aakash_jain@apple.com [Tue, 30 Apr 2019 18:33:30 +0000 (18:33 +0000)]
[ews-build] Parse and display bindings test failures
https://bugs.webkit.org/show_bug.cgi?id=197423

Reviewed by Lucas Forschler.

* BuildSlaveSupport/ews-build/steps.py:
(RunBindingsTests.__init__): Set timeout of 5 minutes.
(RunBindingsTests.start): Initialize log_observer for json output.
(RunBindingsTests.getResultSummary): Update step and build summary based on bindings test results.
(RunBindingsTests._addToLog): Method to add message to log.
* BuildSlaveSupport/ews-build/steps_unittest.py: Updated unit-test accordingly.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244789 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFixing typo in https://trac.webkit.org/changeset/244780/webkit.
sroberts@apple.com [Tue, 30 Apr 2019 18:27:58 +0000 (18:27 +0000)]
Fixing typo in https://trac.webkit.org/changeset/244780/webkit.

Unreviewed test gardening.

* platform/mac/TestExpectations:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244787 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoscrollingcoordinator/scrolling-tree/scrolling-tree-includes-frame.html is a flaky...
sroberts@apple.com [Tue, 30 Apr 2019 18:18:04 +0000 (18:18 +0000)]
scrollingcoordinator/scrolling-tree/scrolling-tree-includes-frame.html is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=197425

Unreviewed test gardening.

* platform/mac-wk2/TestExpectations: Updating expectations for flaky failure

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244786 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFix failing ARM64E wasm tests
keith_miller@apple.com [Tue, 30 Apr 2019 17:57:48 +0000 (17:57 +0000)]
Fix failing ARM64E wasm tests
https://bugs.webkit.org/show_bug.cgi?id=197420

Reviewed by Saam Barati.

This patch fixes a bug in the slow path of our JS->Wasm IC bridge
where we wouldn't untag the link register before tail calling.

Additionally, this patch fixes a broken assert when using setting
Options::useTailCalls=false.

* bytecompiler/BytecodeGenerator.cpp:
(JSC::BytecodeGenerator::emitCallForwardArgumentsInTailPosition):
* wasm/js/WebAssemblyFunction.cpp:
(JSC::WebAssemblyFunction::jsCallEntrypointSlow):

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244783 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoFix the build after r244777.
aestes@apple.com [Tue, 30 Apr 2019 17:51:47 +0000 (17:51 +0000)]
Fix the build after r244777.

* DerivedSources-input.xcfilelist:
* DerivedSources-output.xcfilelist:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244782 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoLong-standing rare crash under -[WKWebView _web_immediateActionAnimationControllerFor...
timothy_horton@apple.com [Tue, 30 Apr 2019 17:37:47 +0000 (17:37 +0000)]
Long-standing rare crash under -[WKWebView _web_immediateActionAnimationControllerForHitTestResultInternal...]
https://bugs.webkit.org/show_bug.cgi?id=197404
<rdar://problem/24589331>

Reviewed by Wenson Hsieh.

* UIProcess/mac/WKImmediateActionController.mm:
(-[WKImmediateActionController _updateImmediateActionItem]):
(-[WKImmediateActionController menuItem:previewItemAtPoint:]):
(-[WKImmediateActionController menuItem:itemFrameForPoint:]):
(-[WKImmediateActionController _animationControllerForDataDetectedLink]):
Add some null checks; _webHitTestResult can be null if we (somehow) get
an immediate action gesture without having previously gotten a
mouseDidMoveOverElement from the Web Content process. Cover all our bases.

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244781 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months agoanimations/play-state-in-shorthand.html is a flaky failure
sroberts@apple.com [Tue, 30 Apr 2019 17:08:54 +0000 (17:08 +0000)]
animations/play-state-in-shorthand.html is a flaky failure
https://bugs.webkit.org/show_bug.cgi?id=192501

Unreviewed test gardening.

* platform/ios-simulator-wk2/TestExpectations: Updating expectations for flaky failure
* platform/mac/TestExpectations: Ditto

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244780 268f45cc-cd09-0410-ab3c-d52691b4dbfc

3 months ago[iOS] Fix crash caused by sandbox violation
pvollan@apple.com [Tue, 30 Apr 2019 16:40:08 +0000 (16:40 +0000)]
[iOS] Fix crash caused by sandbox violation
https://bugs.webkit.org/show_bug.cgi?id=197416
<rdar://problem/50266257>

Reviewed by Brent Fulgham.

Add syscall to the iOS sandbox.

* Resources/SandboxProfiles/ios/com.apple.WebKit.WebContent.sb:

git-svn-id: https://svn.webkit.org/repository/webkit/trunk@244779 268f45cc-cd09-0410-ab3c-d52691b4dbfc